QA Engineer - Manual

ABOUT THE PROJECT We’re building Prometheus: autonomous drone swarms designed for GPS-denied and signal-denied environments—with no pilot, no constant radio link, and no external compute. The goal is simple: when drones stop being tools and start being teammates. Instead of “programming missions”, operators state intent—“Search the second floor.” “Map the west wing and find trapped people.”—and the swarm executes, returns to dock, recharges, and continues. We’re now moving from “it works” to reliability, productization, pilots, and real deployments—that’s the phase where great engineering matters most. Example of our work: THE ROLE You’ll own the next-generation Prometheus drone hardware and the dock. Your job is to take what a small team prototyped and turn it into something we can confidently put into customers’ hands for industrial inspection / indoor operations, with 24/7 capability via docking + self-charging or automatic battery hot-swap. This is a builder role: design → prototype → break → iterate → harden → repeat. Minimal meetings. High ownership. This is not a research-only role. In the early phase, you will also build and assemble drones and docks for customers in low volumes, take responsibility for practical quality checks, and fix things that break in the field. Over time, manufacturing will move to a dedicated process/team, but initially this role includes hands-on building, testing, and pragmatic QA. This role is focused on raising technology readiness (TRL) — turning prototypes into repeatable, reliable hardware suitable for pilots and early customers, not exploratory research. KEY RESPONSIBILITIES: Airframe & endurance * Design the next drone iteration focused on meaningfully longer mission time, stability, robustness, and serviceability (current version V1 can be seen here: https://www.goodai.com/swarm-robotics/ ). * Drive propulsion/weight/thermal tradeoffs with explicit budgets (mass, power, noise, cooling, crash energy, etc.). * Design for close-to-obstacles indoor flight with protections (guards, bump tolerance, maintainability). Docking & energy replenishment (core) * Own the dock mechanical system and the drone-side interface: * precision landing guidance & alignment features * contact charging OR inductive charging OR automatic hot-swap interface (your recommended solution) * safe retention/locking, E-stop behaviors, fail-safe states * Design for repeated cycles, dirt/dust, misalignment, and abuse. Hardening for customer environments * Turn prototype “works in our observation space” into “works reliably for pilots and customers”: * shock/vibration, protective structures, connectors, cable management * repairability: fast swap of arms/guards, modular components * environmental constraints: dust, temperature, light smoke residue, etc. Engineering process (lightweight, but real) * Implement DFM/DFA thinking early (fast assembly, reduced part count, reliable fasteners, tolerances). * Define acceptance tests, reliability tests, and a “flight hours → failures → fixes” loop. * Own suppliers for mechanical parts and iterate rapidly (3D print, CNC, composites—whatever fits). * Comfort defining and tracking basic quality metrics (failure rate, repair time, cycle count). WHAT SUCCESS LOOKS LIKE (FIRST 3–6 MONTHS) * A clear hardware roadmap with 2–3 staged prototypes (alpha → hardened alpha → pilot unit). * A new drone iteration with clearly improved flight time (e.g. via batteries, propellers, airframe efficiency, mass reduction), backed by measurements—not just theory. * A working dock prototype with repeatable autonomous cycles and a realistic path to production. * A documented energy replenishment decision (charge vs swap), backed by a test plan and early validation. * Measurable improvement in endurance, repairability, and “ops friction” (setup time, swap time, failure rate). TRIAL PERIOD AND EVALUATION (FIRST 4 MONTHS) This role has a clear trial period with explicit expectations and fast feedback. * Progress is reviewed weekly directly by Marek. * We prefer to end things early rather than let misalignment drag on. Expected trajectory during the trial period: Weeks 1–4 * Rebuild and iterate on the current V2 drone. * Show fast execution: first working hardware improvements in days, not weeks. * Demonstrate the ability to choose a direction, finish, and not get lost in options. Weeks 5–10 * Deliver a more substantial iteration (informally “V3 direction”): * improved flight time * better robustness and serviceability * cleaner assembly and repair flow Weeks 11–16 * Design and prototype the docking station (charging or hot-swap). * Validate core assumptions with real hardware tests. * Show that you can converge on a reliable solution, not just explore concepts. If progress stalls, iteration speed is low, or ownership is missing, we will not wait months to make a decision. REQUIRED PROFILE (MUST-HAVE) * Strong mechanical/mechatronics engineering skills with real shipped hardware, not only research prototypes. * Excellent 3D modelling tools (CAD, Fusion360 or similar)  * Experience with industrial design  * Hands-on prototyping: 3D printing, workshop tools, fast iteration, debugging physical systems. * Ability to design systems that survive repeated use: connectors, mounts, retention, wear, vibration, shock. * Independence: you can take a vague goal, turn it into requirements + design + test plan, and execute. GOOD TO HAVE * UAV or high-dynamics robotics experience (propulsion, vibration, weight/power budgets). * Docking/charging/hot-swap systems (robots, drones, warehouses, consumer devices). * Basic electronics integration literacy (you don’t need to be an EE, but you can work with one). * Ability to write and debug firmware/software for embedded systems * Experience with CFD, basic understanding of aerodynamics  * Reliability tools: FMEA, HALT/HASS mindset, design of stress tests, failure analysis. * Manufacturing exposure: contract manufacturers, supplier management, assembly line thinking. * Comfort using modern AI tools (e.g. LLMs, AI coding assistants) for faster iteration, documentation, analysis, and problem solving. THIS WILL BE A BAD FIT IF…  * You prefer long planning cycles over building and testing hardware weekly. * You avoid ambiguity and need fully specified requirements before starting. * You don’t enjoy owning systems end-to-end (design → build → test → field failures → fixes). * You’re here mainly for research novelty rather than hardening and shipping. WHAT YOU’LL FIND AT GOODAI: Impact * Build a real-world autonomous drone-swarm platform with measurable field deployments. You’ll own a core subsystem of a category-defining product: LLM-first, swarm-first, full-stack control, reliability-obsessed autonomy. * The tech matters: autonomy in GPS/signal-denied environments is strategically important, and we’re building it in Europe. * Small team, high impact. If you like building real machines fast, you’ll fit. How we work * on-site Prague 6: limited home office * Low bureaucracy, fast decisions, direct access to founder Time off & wellbeing * 5 weeks’ vacation * Flexible sick leave Workspace * Modern offices at Oranžerie, Prague — quiet focus zones, beautiful garden, great daylight, quality equipment, bike storage, gym and shower, fully stocked kitchen, lunches prepared by our private chefs Community * Regular team breakfasts, workshops, and relaxed meetups (BBQ, game nights, parties, teambuildings)