Role Atlas · Mechanical Design / Humanoid Hardware

Mechanical Design Engineer

Mechanical design engineers create the physical parts, assemblies, mechanisms, structures, enclosures, joints, covers, and thermal paths that make a humanoid robot possible.

Plain English:a mechanical design engineer designs the robot's physical body so it can move, survive contact, carry loads, protect electronics, shed heat, and be built repeatedly.

00 · Stack map

Where this role sits in the humanoid stack

  • Body: frames, shells, covers, structures, joint packaging, material selection, impact resistance, service access, thermal paths.
  • Hands: fingers, palms, compliant elements, tactile surfaces, tendons, linkages, compact mechanisms, grip interfaces, sensor and cable packaging.
  • Legs: hip, knee, ankle, foot, and limb structures that balance stiffness, weight, range of motion, durability, and impact loads.
  • Power: battery packaging, cooling, mounting, access, sealing, cable strain relief, charge-port and power-distribution packaging.
  • Simulation layer: CAD-to-analysis, FEA, tolerance stack-ups, thermal models, mass properties, inertia, and simulation-ready robot geometry.
  • Factory layer: drawings, GD&T, DFM/DFA, supplier collaboration, NPI builds, first articles, inspection plans, fixtures, reliability testing, and failure analysis.
01 · The work

What this role actually does

A mechanical design engineer designs the robot's physical hardware so the rest of the stack has a body it can trust.

In a humanoid robotics company, the work often includes:

  • Turning product, system, safety, controls, electrical, and manufacturing requirements into physical concepts.
  • Designing robot parts and assemblies in 3D CAD.
  • Owning subsystems such as hand structures, arm links, torso structures, covers, brackets, joint housings, feet, battery mounts, sensor mounts, cooling ducts, cable carriers, or exterior panels.
  • Creating compact mechanical layouts around actuators, bearings, encoders, sensors, wiring, PCBs, fans, ducts, fasteners, seals, and service features.
  • Balancing trade-offs between mass, stiffness, strength, range of motion, cost, manufacturability, durability, safety, and appearance.
  • Designing mechanisms such as joints, linkages, fingers, palms, compliant elements, latches, covers, doors, service panels, belt paths, tendon paths, or cable routing features.
  • Selecting materials, finishes, coatings, fasteners, bearings, bushings, seals, adhesives, elastomers, polymers, sheet metals, machined parts, castings, or composites.
  • Performing first-order calculations for load, stress, stiffness, fatigue, torque, heat transfer, and tolerance.
  • Running or coordinating FEA, thermal analysis, motion analysis, or other engineering simulations.
  • Creating 2D drawings with clear dimensions, GD&T, notes, datums, finishes, inspection points, and manufacturing requirements.
  • Performing tolerance stack-ups to make sure moving assemblies still work after real parts are manufactured.
  • Building prototypes using 3D printing, CNC machining, sheet metal, laser cutting, waterjet cutting, casting, molding, or vendor parts.
  • Designing test fixtures and validation plans for strength, stiffness, wear, fatigue, drop, impact, thermal, ingress, and lifecycle testing.
  • Working with electrical engineers on sensor placement, PCB packaging, connector access, harness routing, grounding points, and strain relief.
  • Working with controls, locomotion, and manipulation engineers to make sure the hardware supports the needed motion, sensing, compliance, and torque requirements.
  • Working with manufacturing, quality, suppliers, and contract manufacturers to move designs through NPI and into production.
  • Investigating failures found during lab testing, manufacturing, or field deployment and driving corrective actions.
  • Updating CAD, drawings, BOMs, and engineering change orders as the design matures.

This is a hands-on role. The strongest mechanical design engineers do not stop at the CAD model. They build, inspect, test, break, measure, revise, and document.

What the work feels like day to day

A normal week might include:

  • Reworking a hip joint housing because the original design is too heavy and blocks wire harness routing.
  • Reviewing an actuator package with electrical and controls engineers to check bearing loads, encoder placement, heat paths, and assembly sequence.
  • Creating a tolerance stack-up for a finger mechanism that binds when molded parts drift toward worst-case dimensions.
  • Running an FEA study on a robot hand structure after drop-test damage appears near a fastener boss.
  • Changing a bracket from a 3D-printed prototype into a CNC or injection-molded production candidate.
  • Designing a fixture to cycle a joint through thousands of repetitions while measuring looseness, wear, and temperature rise.
  • Inspecting first articles from a supplier and comparing CMM data against drawing requirements.
  • Adding service access so a technician can replace a sensor without removing half the robot.
  • Sitting in a design review and defending a mass/stiffness/manufacturing trade-off with evidence.
  • Updating a BOM and release package after a late supplier change affects material, finish, or tolerances.

The best mechanical design engineers are practical. They understand physics, but they also understand shops, suppliers, technicians, assembly mistakes, cost pressure, tight schedules, and the ugly ways real hardware fails.


02 · Why it matters

Why it matters in humanoid robotics

Humanoid robots are physical AI systems. The software may get attention, but the robot still has to stand, walk, reach, grasp, carry, fall, recover, stay cool, protect electronics, and survive repeated contact with the world.

Mechanical design engineering matters because humanoids need:

  1. A body that matches the intelligence
    A control policy, perception model, or manipulation planner is limited by what the body can physically do. Poor stiffness, backlash, mass distribution, friction, cable routing, or thermal design can make good algorithms look bad.

  2. High performance inside tight packaging
    Humanoid robots try to fit motors, transmissions, sensors, batteries, compute, wiring, cooling, fasteners, structure, and covers into human-like proportions. The packaging problem is severe.

  3. Low mass without fragile hardware
    Every gram matters, especially in limbs and hands. Lower mass improves efficiency, speed, control bandwidth, and safety, but weak or brittle hardware creates failures. Mechanical design is often the fight between light, stiff, durable, and buildable.

  4. Contact-rich durability
    Humanoids touch the world. Hands grip objects, feet strike floors, arms bump shelves, covers take impacts, and joints cycle constantly. Mechanical designs must survive repeated contact, not just static loads.

  5. Thermal and energy discipline
    Actuators, batteries, compute, and power electronics generate heat inside a dense moving body. A beautiful mechanism that overheats quickly is not deployable.

  6. Manufacturability at scale
    A prototype can be hand-fitted by an expert. A production robot cannot. Mechanical design engineers make parts that can be manufactured, inspected, assembled, serviced, and improved.

  7. Safety around people
    Human-facing robot hardware needs smooth edges, controlled pinch points, safe covers, compliant interfaces where appropriate, predictable failure modes, and repairable structures.

  8. A bridge between hardware and every other team
    Mechanical design touches electrical packaging, embedded systems, controls, perception, manipulation, locomotion, manufacturing, product, safety, and field operations. A mechanical decision can help or hurt every layer of the stack.

A simple rule: if the robot's body is too heavy, hot, fragile, hard to assemble, or hard to repair, the company cannot scale no matter how good the AI demo looks.


03 · Backgrounds

Best-fit backgrounds

This role is a strong fit for people who like physical systems, precise design, hands-on building, and cross-functional trade-offs.

Mechanical engineers from robotics, mechatronics, or automation

You already have useful skills: CAD, machine design, mechanisms, motors, sensors, prototyping, basic controls awareness, and hands-on integration.

You are probably missing: humanoid-specific packaging constraints, high-cycle wearable-like mechanisms, dexterous hands, lightweight structures, robot-scale thermal constraints, and production discipline at higher volumes.

Best entry angle: robotics mechanical engineer, mechanical design engineer, mechanisms engineer, hand structures engineer, robot body engineer, mechanical systems engineer.

Mechanical engineers from consumer electronics or product design

You already have useful skills: compact packaging, plastics, surfacing, high-volume manufacturing, design for assembly, tooling, CM/vendor collaboration, cosmetic quality, and tolerance discipline.

You are probably missing: actuators, bearings, torque paths, kinematics, dynamic loads, robot falls, sensor/actuator integration, and continuous motion lifecycle testing.

Best entry angle: robot exteriors, covers, hands, tactile surfaces, enclosures, battery packaging, electronics packaging, compact mechanism design, production hardware design.

Automotive, EV, aerospace, or medical-device engineers

You already have useful skills: structural design, materials, reliability, testing, documentation, DFMEA, tolerance, suppliers, safety, thermal, and production quality.

You are probably missing: fast prototype cycles, humanoid proportions, high-density actuation, dexterous manipulation, and integration with robot learning, controls, and autonomy teams.

Best entry angle: structures, mechanisms, battery packaging, thermal systems, actuator housings, joints, safety-critical hardware, mechanical validation, reliability engineering.

Mechanical engineering, robotics, or mechatronics students

You may already have coursework in CAD, statics, dynamics, materials, manufacturing, robotics, machine design, or mechatronics.

You are probably missing: professional CAD discipline, GD&T, tolerance stack-ups, supplier-ready drawings, design reviews, validation reports, failure analysis, and evidence that your parts can be built and tested.

Best entry angle: mechanical engineering intern, robotics hardware intern, junior mechanical design engineer, mechanical test intern, design-for-manufacturing intern.

CAD designers and mechanical drafters

You already have useful skills: clean CAD, drawings, model organization, BOMs, revisions, and design documentation.

You are probably missing: engineering ownership, requirements trade-offs, load calculations, material selection, test planning, tolerance reasoning, and design validation.

Best entry angle: CAD designer for robotics, mechanical designer, junior mechanical engineer, documentation-heavy hardware roles, then grow toward full subsystem ownership.

Manufacturing, NPI, quality, or mechanical test engineers

You already understand how hardware fails in production: assembly variation, inspection problems, supplier escapes, fixture limitations, test repeatability, and process constraints.

You are probably missing: early concept generation, architecture trade-offs, CAD ownership, design calculations, and mechanism design.

Best entry angle: production-minded mechanical design, DFM/DFA-heavy roles, mechanical validation, supplier development, reliability-to-design roles.

Industrial designers moving toward robotics hardware

You already understand human-facing form, ergonomics, surface quality, user perception, and product coherence.

You are probably missing: structural load paths, mechanisms, tolerances, materials engineering, manufacturing constraints, and electromechanical integration.

Best entry angle: robot exteriors, covers, human-facing interfaces, softgoods, service-access design, then deepen mechanical engineering fundamentals.


04 · Skills

Skills to learn

Think of this role in layers: engineering fundamentals, CAD discipline, mechanisms, analysis, manufacturing, test, and humanoid-specific integration.

Core mechanical engineering fundamentals

These are the foundation for credible mechanical design work.

  • Statics and dynamics.
  • Strength of materials: stress, strain, stiffness, deflection, fatigue, buckling, stress concentration.
  • Machine design: fasteners, bearings, shafts, springs, gears, belts, pulleys, cams, linkages, bushings, pins, keys, retaining rings, adhesives, and fits.
  • Materials: aluminum, steel, titanium, magnesium, engineering plastics, elastomers, composites, foams, coatings, surface treatments, and thermal interface materials.
  • Kinematics: joints, linkages, degrees of freedom, constraints, range of motion, singularities at a practical level.
  • Friction, wear, lubrication, backlash, compliance, and hysteresis.
  • Thermal basics: conduction, convection, radiation, heat spreading, airflow, fan/duct trade-offs, and duty cycles.
  • First-order calculations before simulation.

CAD and design intent

Robot hardware teams need clean, editable CAD, not fragile models.

Learn:

  • Parametric CAD modeling.
  • Large assembly management.
  • Top-down and bottom-up design methods.
  • Skeleton models, master geometry, datums, coordinate systems, and design intent.
  • Surfacing for covers, hands, exterior panels, and human-facing parts.
  • Configuration management and revision control for mechanical design.
  • Clearance studies, interference checks, range-of-motion checks, and service-access checks.
  • Mass properties, center of mass, inertia, and export to simulation.
  • Drawing creation with clear manufacturing and inspection intent.
  • BOM management and release discipline.

GD&T, tolerances, and drawings

Robotics hardware must move after real parts are manufactured.

Learn:

  • ASME Y14.5 or ISO GPS basics.
  • Datums, feature control frames, profile, position, runout, flatness, perpendicularity, parallelism, concentricity, and cylindricity.
  • Fits and allowances for bearings, shafts, pins, bushings, housings, and sliding interfaces.
  • Tolerance stack-up methods.
  • Inspection planning and CMM-friendly design.
  • Critical-to-quality dimensions.
  • How manufacturing processes affect achievable tolerances.
  • How to write drawings that suppliers can actually build from.

Mechanisms and robot-specific hardware

Humanoids need mechanisms that are compact, durable, and easy to control.

Learn:

  • Joint design for hips, knees, ankles, shoulders, elbows, wrists, fingers, and necks.
  • Bearing selection, load paths, preload, alignment, and service life.
  • Gear trains, belts, tendons, cable drives, harmonic drives, cycloidal drives, planetary gearboxes, ball screws, lead screws, and linkages.
  • Underactuated and compliant mechanisms.
  • Mechanical stops and range-of-motion limits.
  • Routing for cables, tendons, tubes, sensors, and flexible circuits.
  • Design for backlash, stiffness, friction, noise, and wear.
  • Packaging of motors, encoders, brakes, sensors, motor drivers, PCBs, and connectors.
  • Safe covers and pinch-point reduction.

Humanoid-specific design constraints

These make humanoid hardware different from many normal machines.

  • Human-like proportions and tight body envelopes.
  • High joint density and limited service volume.
  • Mass distribution that affects balance and locomotion.
  • Limb inertia that affects control performance and energy use.
  • Impact loads from walking, falls, collisions, and repeated contact.
  • Hands that need strength, compliance, sensing, durability, and dexterity.
  • Feet that need traction, structure, sensing, and replaceable wear surfaces.
  • Shoulders, hips, and wrists that need wide range of motion without cable failure.
  • Exterior surfaces that must protect the robot and be safe around people.
  • Thermal behavior under dynamic duty cycles, not just steady-state operation.

Analysis and simulation

Simulation is useful, but only when the model matches the physical question.

Learn:

  • Hand calculations for sanity checks.
  • Finite element analysis for stress, strain, stiffness, fatigue, and modal behavior.
  • Contact modeling basics and the limits of simulation.
  • Thermal modeling and CFD basics.
  • Motion analysis and multibody dynamics at a practical level.
  • Correlating simulation with physical test data.
  • Sensitivity studies and design of experiments.
  • Failure margins, safety factors, and uncertainty.
  • When not to over-trust a beautiful simulation plot.

Manufacturing and production readiness

A humanoid robot is not useful if it can only be assembled by the original designer.

Learn:

  • DFM and DFA.
  • CNC machining, sheet metal, casting, die casting, injection molding, compression molding, overmolding, additive manufacturing, composites, bonding, and fastening.
  • Tooling concepts: draft, ribs, bosses, wall thickness, gates, ejector pins, slides, lifters, parting lines, inserts, and mold flow.
  • Supplier communication and design reviews.
  • First article inspection.
  • NPI builds and build readiness reviews.
  • Assembly sequence design.
  • Error-proofing and serviceability.
  • Cost drivers in mechanical parts.
  • Engineering change orders and release control.

Testing, validation, and failure analysis

Mechanical design becomes credible when it survives tests.

Learn:

  • Test plan creation.
  • Instrumentation: load cells, force gauges, torque sensors, strain gauges, thermal cameras, thermocouples, accelerometers, displacement sensors, DAQ systems.
  • Static load testing.
  • Cycle-life testing.
  • Impact, drop, vibration, and shock testing.
  • Thermal testing.
  • Wear testing.
  • Environmental testing: dust, moisture, temperature, sealing, sweat/oil/chemical exposure where relevant.
  • Root-cause analysis, 5 Whys, fishbone diagrams, fault trees, and FA/CA reports.
  • Turning failures into design changes instead of excuses.

05 · Tools

Tools & technologies

Do not present this list as a syllabus where every tool is required. Different companies use different stacks. These are the common clusters to recognize.

CAD and mechanical design

  • CATIA / 3DEXPERIENCE: common in automotive, aerospace, complex surfacing, and large assembly environments.
  • Siemens NX: common in advanced manufacturing, robotics, aerospace, and complex mechanical systems.
  • SolidWorks: common in robotics startups, product design, research labs, prototypes, and supplier communication.
  • PTC Creo: common in hardware companies, production engineering, and complex mechanical design.
  • Onshape: useful for cloud CAD, collaboration, version control, and fast distributed hardware work.
  • Fusion 360: useful for personal projects, prototypes, CAM workflows, and early portfolio work.

PDM, PLM, BOM, and release systems

  • Teamcenter.
  • Windchill.
  • ENOVIA / 3DEXPERIENCE.
  • Arena PLM.
  • Onshape built-in PDM.
  • Propel, Omnify, or other startup-friendly PLM systems.
  • Jira, Linear, or issue trackers for design tasks and engineering changes.
  • ECO / ECR workflows.
  • BOM management and revision control.

Analysis and simulation

  • ANSYS Mechanical: structural FEA, fatigue, modal, contact, and thermal-mechanical analysis.
  • ANSYS Fluent / Icepak: CFD and electronics/thermal work.
  • Abaqus: nonlinear materials, contact, fatigue, and complex simulation.
  • MSC Nastran / Patran / FEMAP: structural analysis and aerospace-style workflows.
  • SolidWorks Simulation: accessible FEA integrated into SolidWorks.
  • Simcenter / NX CAE: integrated simulation workflows.
  • MATLAB / Python: calculations, data analysis, test processing, tolerance studies, and plotting.
  • Excel / Google Sheets: first-order calculations and tolerance stack-ups, when used carefully.

Manufacturing and prototyping

  • 3D printers: FDM, SLA, SLS, MJF, DMLS where available.
  • CNC mills and lathes.
  • Sheet metal tools.
  • Laser cutters and waterjet cutters.
  • Injection molding and mold-flow review tools.
  • Casting, die casting, and forging supplier processes.
  • Softgoods, elastomers, overmolding, and compression molding workflows.
  • Surface finishing, coating, anodizing, plating, painting, and texturing.

Drawings, inspection, and quality

  • GD&T standards such as ASME Y14.5 or ISO GPS.
  • CMM inspection.
  • Optical scanning and structured-light scanning.
  • Calipers, micrometers, height gauges, gauge pins, torque tools, and inspection fixtures.
  • Cp/Cpk and basic statistical process control.
  • Design of experiments.
  • First article inspection reports.
  • Supplier quality documentation.

Robotics integration tools

Mechanical design engineers do not need to be robotics software experts, but these tools help communication with the rest of the team:

  • URDF / xacro for robot description.
  • SDF where simulation tools require it.
  • ROS 2 awareness for coordinate frames, sensor placement, and mechanical/electrical/software interfaces.
  • Isaac Sim, Gazebo, MuJoCo, or other simulation environments for geometry, mass, inertia, and clearance validation.
  • CAD export workflows for simulation and collision geometry.
  • Robot log review tools when mechanical failures appear during motion tests.

Lab and test equipment

  • Torque wrenches and torque drivers.
  • Force gauges and load cells.
  • Strain gauges.
  • Accelerometers.
  • Thermocouples and thermal cameras.
  • Environmental chambers.
  • Shaker tables where available.
  • Instron or universal testing machines.
  • High-speed cameras.
  • DAQ systems.
  • Basic electronics tools for electromechanical debugging: multimeter, oscilloscope, power supply, and continuity checks.

06 · Projects

Portfolio projects to prove ability

A strong mechanical design portfolio should show more than renderings. It should show requirements, CAD, trade-offs, calculations, prototypes, tests, failures, revisions, drawings, and manufacturing thinking.

The goal is not to build a complete humanoid robot. The goal is to show that you can design robot hardware like an engineer, not only like a maker.

Project 1: Compact robotic joint module

Build: a CAD and prototype project for a compact rotary joint that includes a motor placeholder or real motor, gearbox or belt/tendon transmission, bearings, encoder placement, hard stops, cable routing, covers, and mounting interfaces.

Use affordable components if needed. The point is not the most powerful actuator. The point is good mechanical reasoning.

What it proves:

  • You understand joint packaging.
  • You can design load paths and bearing support.
  • You can manage constraints around motors, wires, sensors, fasteners, and service access.
  • You can reason about range of motion, backlash, stiffness, and assembly sequence.
  • You can turn a concept into CAD, drawings, and a tested prototype.

Evidence to include:

  • Requirements table.
  • Exploded CAD view.
  • Section views showing load path and packaging.
  • Mass and estimated inertia.
  • Basic torque and bearing-load calculations.
  • FEA or hand calculations for critical parts.
  • Drawings with tolerances and material/finish notes.
  • Photos or video of prototype assembly.
  • Test results and design revisions.

Project 2: Lightweight robot limb or structural link

Build: a robot arm, leg, or torso structural link optimized for stiffness, mass, durability, and manufacturability. Compare at least two design concepts and explain the trade-offs.

This can be a forearm link, thigh link, foot structure, torso bracket, shoulder frame, or simplified humanoid limb section.

What it proves:

  • You can reason about load paths.
  • You can balance mass and stiffness.
  • You can use material selection and geometry intelligently.
  • You can validate a structure with calculations and testing.
  • You can communicate design decisions clearly.

Evidence to include:

  • Load case assumptions.
  • Concept comparison.
  • CAD model and drawings.
  • FEA screenshots with boundary conditions explained.
  • Physical test setup if built.
  • Measured deflection or failure load.
  • Notes on how the part would change for production.

Project 3: Dexterous end-effector or robotic finger mechanism

Build: a small robotic finger, gripper, or palm module that demonstrates mechanism design, compliance, grasp contact, cable routing, and compact packaging.

Keep the scope realistic. A single well-documented finger is better than a five-finger hand that barely works.

What it proves:

  • You understand contact-rich mechanism design.
  • You can package tendons, linkages, springs, bearings, pins, and sensors.
  • You can design for grip, compliance, and durability.
  • You can think about tactile surfaces, friction, and wear.
  • You understand why hands are mechanically difficult.

Evidence to include:

  • Kinematic diagram.
  • CAD model and motion study.
  • BOM.
  • Prototype photos.
  • Grasp tests with different objects.
  • Notes on failure modes such as tendon wear, binding, backlash, or broken prints.
  • Redesign notes after testing.

Project 4: Thermal and electronics packaging module

Build: an enclosure or robot-body module that packages a compute board, motor driver, battery cell placeholder, fan, heat sink, duct, sensor, and connectors under tight space constraints.

Show how heat leaves the system and how a technician would assemble and service it.

What it proves:

  • You understand electromechanical packaging.
  • You can design around heat, cables, connectors, airflow, service access, and mounting.
  • You can communicate with electrical and embedded teams.
  • You can reason about duty cycles and thermal paths.

Evidence to include:

  • Packaging layout.
  • Airflow or heat-path diagram.
  • First-order thermal calculation.
  • CFD or simple thermal simulation if available.
  • Assembly sequence.
  • Cable strain-relief details.
  • Serviceability notes.
  • Test plan for measuring temperature rise.

Project 5: Prototype-to-production redesign

Build: take a rough 3D-printed robot part and redesign it as a production-intent part for CNC, sheet metal, casting, injection molding, or another realistic process.

Do not only show the final model. Show what changed and why.

What it proves:

  • You can move beyond maker prototypes.
  • You understand DFM/DFA.
  • You can create supplier-ready drawings.
  • You can think about tolerances, inspection, tooling, cost, and assembly.
  • You can communicate manufacturing intent.

Evidence to include:

  • Before/after CAD.
  • Manufacturing-process selection.
  • Material selection.
  • Drawing with GD&T.
  • Tolerance stack-up.
  • Assembly steps.
  • Inspection plan.
  • Cost or process trade-off notes.

Project 6: Mechanical validation and failure-analysis report

Build: a test fixture and report for a mechanical component. Test static load, repeated cycles, impact, heat, wear, or looseness. Then identify a failure and propose a design correction.

This can be done with simple hardware if the test is well documented.

What it proves:

  • You understand validation.
  • You can measure physical performance.
  • You can turn failure into engineering action.
  • You can write the kind of report hardware teams actually need.

Evidence to include:

  • Test objective and acceptance criteria.
  • Fixture CAD or photos.
  • Instrumentation setup.
  • Raw data and plots.
  • Failure photos.
  • Root-cause analysis.
  • Corrective action and revised design.
  • Retest plan.

07 · Titles

Common job titles

Mechanical design jobs in humanoid robotics use many titles. Use these titles and keywords when building the jobs taxonomy.

Direct titles

  • Mechanical Design Engineer
  • Robotics Mechanical Engineer
  • Mechanical Engineer, Robotics
  • Mechanical Hardware Engineer
  • Mechanical Systems Engineer
  • Product Design Engineer, Robotics
  • Senior Mechanical Engineer
  • Staff Mechanical Engineer
  • Principal Mechanical Engineer
  • Mechanical Engineer, Humanoid Robot

Specialist titles

  • Mechanical Engineer, Structures
  • Mechanical Engineer, Robot Body
  • Mechanical Engineer, Hands
  • Mechanical Engineer, Dexterous Manipulation
  • Mechanical Engineer, End Effector
  • Mechanical Engineer, Mechanisms
  • Mechanical Engineer, Joints
  • Mechanical Engineer, Exteriors
  • Mechanical Engineer, Softgoods
  • Mechanical Engineer, Elastomers
  • Mechanical Engineer, Injection Molding
  • Mechanical Engineer, Battery Packaging
  • Mechanical Engineer, Thermal Systems
  • Mechanical Engineer, Interconnects
  • Mechanical Packaging Engineer
  • Actuator Mechanical Design Engineer
  • Robotics CAE Engineer
  • Mechanical Analysis Engineer
  • Mechanical Test and Validation Engineer
  • Mechanical Reliability Engineer

Adjacent titles

  • Mechatronics Engineer
  • Hardware Engineer, Robotics
  • Systems Integration Engineer, Hardware
  • NPI Mechanical Engineer
  • Supplier Development Engineer, Mechanical
  • Manufacturing Engineer, Mechanical
  • Quality Engineer, Mechanical Hardware
  • Design Release Engineer
  • Mechanical Engineering Intern, Robotics

Search keywords

Use these as job-board filters:

  • mechanical design engineer robotics
  • robotics mechanical engineer
  • humanoid mechanical engineer
  • robot hardware engineer
  • robot structures engineer
  • robot mechanisms engineer
  • end effector mechanical engineer
  • dexterous hand mechanical engineer
  • mechanical engineer hands
  • actuator mechanical design
  • robot body design
  • robot exteriors
  • thermal mechanical engineer robotics
  • battery packaging robotics
  • GD&T robotics
  • tolerance stack-up robotics
  • DFMA robotics
  • mechanical validation robotics
  • FEA robotics
  • CAD robotics
  • CATIA robotics
  • NX robotics
  • SolidWorks robotics
  • Creo robotics

08 · Companies

Companies hiring for this work

Job openings change quickly. Treat this as a live company map, not a permanent list. These are strong examples to seed the Companies and Jobs sections.

Figure

Figure hires mechanical engineers across humanoid hands, structures, compliant elements, battery packaging, thermal prototyping, interconnects, and robot ecosystem hardware. Current examples reviewed on 2026-07-02 included mechanical design roles requiring 3D CAD, GD&T to ASME Y14.5, tolerance stack-ups, FEA, prototyping, validation, NPI support, and failure analysis.

Why it matters for this role: Figure's listings are good examples of mechanical design work that moves from concept through production. They show that humanoid mechanical engineering includes hands, structures, softgoods, plastics, thermal systems, batteries, sensors, wiring, suppliers, inspection, and field failure analysis.

Useful internal links to create:

  • /careers/companies/figure
  • /careers/jobs?company=figure&role_family=mechanical-design
  • /careers/role-atlas/manipulation-engineer
  • /careers/role-atlas/actuator-engineer
  • /careers/role-atlas/electrical-systems-engineer
  • /careers/role-atlas/manufacturing-engineer

Apptronik

Apptronik hires mechanical engineers for Apollo across staff mechanical engineering, R&D, end-effector structures, actuation, structural simulation, thermal simulation, upper-body design, and manufacturing-facing hardware roles.

Why it matters for this role: Apptronik's listings are useful because they make the full-lifecycle nature of mechanical design explicit: concept, architecture, design trade studies, NPI, production release, sustaining, reliability testing, root-cause analysis, and cross-functional work with electrical, software, manufacturing, quality, supply chain, and operations.

Useful internal links to create:

  • /careers/companies/apptronik
  • /careers/role-atlas/actuator-engineer
  • /careers/role-atlas/manipulation-engineer
  • /careers/role-atlas/simulation-engineer
  • /careers/role-atlas/manufacturing-engineer

Tesla Optimus

Tesla hires for Optimus mechanical hardware under titles such as robotics mechanical design engineer, structures and mechanisms engineer, hand engineer, tactile/sensing packaging engineer, soft goods and exteriors engineer, actuator mechanical design engineer, mechanical test and validation engineer, and robotics CAE engineer.

Why it matters for this role: Optimus roles show the breadth of mechanical design inside a humanoid program: structures, mechanisms, hands, tactile sensing, exteriors, actuation, gears, validation, and production-minded engineering.

Useful internal links to create:

  • /careers/companies/tesla-optimus
  • /careers/role-atlas/actuator-engineer
  • /careers/role-atlas/robot-test-validation-engineer
  • /careers/role-atlas/electrical-systems-engineer
  • /careers/role-atlas/robotics-ai-engineer

1X Technologies

1X hires across hardware engineering, manufacturing operations, robot safety, materials, actuators, batteries, hands, test, and mechanical engineering roles around NEO and related robot systems.

Why it matters for this role: 1X is useful for candidates interested in home-facing humanoids, mechanical design around hands and bodies, battery and material systems, test, validation, safety, service, and manufacturing scale-up.

Useful internal links to create:

  • /careers/companies/1x-technologies
  • /careers/role-atlas/actuator-engineer
  • /careers/role-atlas/robot-test-validation-engineer
  • /careers/role-atlas/robotics-safety-engineer
  • /careers/role-atlas/robot-operations-fleet-operator

Agility Robotics

Agility Robotics builds Digit for industrial work and hires across engineering, manufacturing, AI, safety, quality, and robot deployment functions depending on hiring cycle.

Why it matters for this role: Agility is a useful example for mechanical designers interested in commercial-scale mobile humanoid hardware. Digit-like systems place heavy emphasis on durability, repetitive loaded tasks, maintenance, high reliability, and manufacturable robot bodies.

Useful internal links to create:

  • /careers/companies/agility-robotics
  • /careers/role-atlas/locomotion-engineer
  • /careers/role-atlas/robot-test-validation-engineer
  • /careers/role-atlas/manufacturing-engineer
  • /careers/role-atlas/field-robotics-engineer

Boston Dynamics

Boston Dynamics hires mechanical engineers and roboticists across advanced mobile robots, including Atlas-related mechanical hardware, structures, mechanisms, grippers, robot applications, testing, and platform engineering.

Why it matters for this role: Boston Dynamics is one of the clearest examples of high-performance robot mechanical design. Its Atlas-related hardware work is useful for candidates interested in structures, arms, legs, motors, batteries, grippers, mobility, durability, and real robot testing.

Useful internal links to create:

  • /careers/companies/boston-dynamics
  • /careers/role-atlas/locomotion-engineer
  • /careers/role-atlas/manipulation-engineer
  • /careers/role-atlas/robot-test-validation-engineer
  • /careers/role-atlas/robotics-technical-program-manager

Sanctuary AI

Sanctuary AI hires across AI, robotics, software, controls, sensing, electronics, mechanical engineering, and humanoid systems depending on hiring cycle.

Why it matters for this role: Sanctuary is especially relevant for mechanical designers interested in dexterous humanoid hardware, robotic hands, tactile interaction, physical AI, and mechatronic systems that combine sensing, actuation, and control.

Useful internal links to create:

  • /careers/companies/sanctuary-ai
  • /careers/role-atlas/manipulation-engineer
  • /careers/role-atlas/actuator-engineer
  • /careers/role-atlas/robotics-ai-engineer

NEURA Robotics

NEURA Robotics hires hardware engineers for humanoid robots, including mechanical engineer, injection molding, advanced materials, technical drafting, hardware architecture, mechanical operations, and related roles.

Why it matters for this role: NEURA's humanoid mechanical roles are useful examples of classic mechanical design requirements in humanoids: joints, structures, gearboxes, actuators, CAD, technical drawings, strength analysis, motion analysis, tolerance analysis, material selection, manufacturing process selection, and prototype integration.

Useful internal links to create:

  • /careers/companies/neura-robotics
  • /careers/role-atlas/actuator-engineer
  • /careers/role-atlas/simulation-engineer
  • /careers/role-atlas/electrical-systems-engineer
  • /careers/role-atlas/manufacturing-engineer

09 · Interview

Interview signals

A candidate becomes credible for mechanical design roles when they can show evidence in these areas.

Strong positive signals

  • Can explain a mechanical design from requirements through CAD, prototype, test, failure, revision, and production intent.
  • Shows real CAD, not only renders.
  • Has drawings with meaningful dimensions, datums, tolerances, materials, finishes, and inspection intent.
  • Understands load paths, stiffness, fatigue, wear, backlash, thermal paths, and manufacturing variation.
  • Can explain why a part is shaped the way it is.
  • Can do first-order calculations before opening an FEA tool.
  • Uses FEA or simulation with clear boundary conditions and honest limitations.
  • Has built and tested physical prototypes.
  • Can show failure photos and explain what changed afterward.
  • Understands DFM/DFA and supplier communication.
  • Can discuss tolerance stack-ups and how variation affects robot motion.
  • Understands electromechanical packaging: sensors, motors, wires, connectors, PCBs, heat, and service access.
  • Can communicate trade-offs across mechanical, electrical, controls, manufacturing, and product teams.
  • Has a portfolio that includes test evidence, not only final images.

Weak signals

  • Only shows polished renders with no drawings, calculations, prototypes, or tests.
  • Treats CAD as art instead of engineering.
  • Cannot explain load cases.
  • Cannot explain why a material or process was chosen.
  • Uses FEA screenshots without explaining constraints, loads, mesh, assumptions, or correlation.
  • Has no tolerance reasoning.
  • Designs parts that cannot be assembled, inspected, serviced, or manufactured.
  • Ignores cables, connectors, fasteners, access, cooling, and tooling.
  • Overuses 3D printing as the final answer without discussing production alternatives.
  • Cannot explain a failure or design revision.
  • Blames manufacturing or suppliers without showing design intent clearly.

Interview questions to prepare for

  • Walk me through a mechanical system you designed from concept to test.
  • What were the requirements, and which trade-offs mattered most?
  • How did you choose the material and manufacturing process?
  • What load cases did you design for?
  • How did you validate the design?
  • What failed during testing, and what did you change?
  • How would you reduce mass without losing stiffness or durability?
  • How would you package a motor, gearbox, encoder, bearing, cable harness, and cover inside a small robot joint?
  • How would you design a robotic hand finger for repeated contact and wear?
  • How do you decide between CNC machining, injection molding, casting, sheet metal, and additive manufacturing?
  • Explain a tolerance stack-up you have done.
  • How do you write a drawing so a supplier understands the design intent?
  • How would you make a robot cover safe, serviceable, and manufacturable?
  • How do you work with electrical engineers when PCB placement and mechanical packaging conflict?
  • What makes a mechanical design production-ready?

10 · Pitfalls

Mistakes to avoid

  • Only showing renders. Renders are not enough. Show requirements, CAD, drawings, calculations, test evidence, and revisions.
  • Treating 3D printing as production. 3D printing is useful for prototypes, but production hardware needs process-specific design thinking.
  • Skipping GD&T. Robot parts need clear manufacturing and inspection intent. Learn the basics.
  • Ignoring tolerance stack-ups. Moving assemblies often fail because real parts are not perfect.
  • Over-trusting simulation. FEA is only as good as the assumptions, loads, boundary conditions, material data, and correlation.
  • Forgetting wires and service access. Many robot designs fail because nobody left room for harnesses, connectors, tools, or technicians.
  • Designing for the demo only. Humanoid hardware needs cycle life, impact resistance, repairability, and production readiness.
  • Ignoring controls and software. Mass, inertia, stiffness, backlash, friction, and sensor placement all affect robot behavior.
  • Making everything too strong and heavy. Overbuilding can damage locomotion, energy use, agility, and safety.
  • Making everything too light and fragile. Humanoid robots hit the world. Strength and durability matter.
  • Not documenting failures. A good failure analysis is often more impressive than a perfect-looking final design.

11 · Plan

30 / 60 / 90-day learning plan

This section is optional on Role Atlas pages, but useful for readers who are ready to act.

First 30 days: build the mechanical base

  • Pick one target subsystem: joint, finger, limb link, cover, battery package, or thermal enclosure.
  • Refresh statics, strength of materials, bearings, fasteners, and basic mechanism design.
  • Build a clean CAD model with clear datums, named features, and editable design intent.
  • Learn or refresh GD&T basics.
  • Create first-order calculations for loads, torque, stiffness, or heat.
  • Document assumptions in plain English.

Output: a CAD model, requirements table, calculation sheet, and first draft drawing for a small robot-relevant subsystem.

Days 31–60: prototype, analyze, and test

  • Build or fabricate a prototype.
  • Run a simple FEA, motion study, thermal study, or tolerance stack-up.
  • Create a test plan with acceptance criteria.
  • Measure something physical: force, deflection, temperature, wear, looseness, or cycle count.
  • Photograph failures and update the design.
  • Create a cleaner drawing and BOM.

Output: a prototype, test report, FEA or analysis notes, failure analysis, and revised CAD.

Days 61–90: make it look hireable

  • Turn the project into a professional portfolio case study.
  • Add exploded views, section views, load-path diagrams, and assembly sequence.
  • Add a prototype-to-production redesign section.
  • Add supplier/manufacturing notes.
  • Add tolerance reasoning and inspection intent.
  • Record a short video walkthrough.
  • Map the project to real job descriptions and keywords.

Output: a portfolio page that shows engineering judgment from requirements through prototype, validation, and production intent.


12 · FAQ

FAQ

Is Mechanical Design Engineer a good entry role for humanoid robotics?

Yes, especially for mechanical engineering, mechatronics, product design, automotive, aerospace, consumer electronics, and hardware prototyping backgrounds. It is not an easy shortcut, though. You need evidence that you can design real parts, build prototypes, write drawings, reason about tolerances, and validate hardware.

Do I need to know robotics software?

You do not need to be a robotics software engineer, but you should understand enough to work with software, controls, and electrical teams. Learn how mass, inertia, stiffness, backlash, sensor placement, cable routing, and thermal behavior affect robot performance.

Which CAD tool should I learn?

Learn one serious parametric CAD tool well. SolidWorks, NX, CATIA, Creo, and Onshape are all useful depending on the company. The deeper skill is not the button layout. The deeper skill is design intent, assemblies, drawings, tolerances, release discipline, and manufacturable geometry.

Do I need FEA?

You should understand FEA enough to use it responsibly or communicate with an analysis engineer. You still need hand calculations and physical tests. A clean FEA plot without good assumptions is weak evidence.

Is this role different from Actuator Engineer?

Yes. Mechanical design engineers may design actuator housings, mounts, structures, or packaging, but Actuator Engineer goes deeper into motors, transmissions, torque density, bearings, thermal behavior, motor controls interfaces, and actuator performance.

What is the fastest credible portfolio project?

A compact robotic joint or finger module with CAD, drawings, tolerance notes, calculations, prototype photos, test results, and one documented failure/revision is stronger than a full humanoid concept with no engineering evidence.

Are there non-PhD mechanical roles in humanoid robotics?

Yes. Most mechanical design roles are engineering roles, not research-only roles. Companies need people who can design, build, test, release, manufacture, and fix hardware. Advanced degrees can help for specialized analysis or research, but a strong hardware portfolio can be very powerful.

How much manufacturing should I know?

Enough to avoid designs that cannot be built. Learn the basics of CNC machining, sheet metal, injection molding, additive manufacturing, casting, fastening, assembly, inspection, and supplier communication. Production awareness is a major differentiator.

Take the next step

Ready to apply?

We track every humanoid robotics role from every company building physical AI. Filter by function, vendor, and location.