Software Engineer - Instawork Robotics Labs

Instawork is on a mission to create meaningful economic opportunities for skilled hourly professionals in communities around the globe. Our AI-powered labor marketplace helps local businesses scale, and enables global technology companies to push the frontiers of robotics and AI. Backed by world-class investors like Benchmark, Spark Capital, Craft Ventures, Greylock, Y Combinator, and others, we’re looking for exceptional talent to reimagine the way the world works.

About IRL (Instawork Robotics Labs)

Researchers at UC Berkeley have identified a “100,000-year data gap” — the gulf between what trained AI language models and what physical robots actually have to learn from. Closing that gap is the defining infrastructure challenge of the physical AI era. IRL is Instawork’s answer to it. We deploy skilled workers into real commercial and residential environments — kitchens, warehouses, hotel floors, homes — to capture the high-fidelity task data that the world’s leading Robotics labs use to train their foundation models. Read more [here]

The Role

We are looking for an early-career Software Engineer to join our Bengaluru robotics team and own day-to-day debugging, bring-up, and reliability for InstaCore — our custom multi-camera / sensor hardware used in physical AI data collection. This role sits at the intersection of software and hardware: you will primarily write and reason about software (scripts, diagnostics, firmware interfaces, calibration tooling), while being comfortable opening a device, tracing a failure from log → sensor → board, and driving it to a fix.

This is an AI-first engineering role. A large share of the work — writing diagnostics, parsing logs, drafting RCA notes, exploring unfamiliar codebases, and spinning up small tools — can be accelerated dramatically with modern coding agents. We strongly prefer people who already use tools like Cursor, Claude Code, and Codex heavily in their day-to-day work, and who treat AI as a default collaborator rather than an occasional shortcut.

Who You Are:

• An early-career engineer, from a recent graduate up to ~2 years of experience, with a Bachelor’s (or equivalent) in Computer Science, Electrical / Electronics, Mechatronics, Mechanical, or a related engineering field.
• AI-first by default. You regularly use coding agents and AI IDEs (especially Cursor, Claude Code, and/or Codex) to write, debug, and ship faster. You know how to prompt well, verify outputs, and keep ownership of correctness — you don’t blindly paste.
• Software-first, systems-curious. You are strongest writing and debugging code (Python and/or C/C++), and you enjoy chasing bugs through software into the physical world.
• A hands-on debugger. You are comfortable with Linux shells, reading logs, flashing firmware, checking cables/power, and using basic lab tools (multimeter, serial console) when software alone isn’t enough.
• Comfortable with embedded and robotics concepts. Exposure to microcontrollers (e.g. ESP32, STM32), cameras/sensors, UART/I2C/SPI/CAN, or ROS is a plus — curiosity and follow-through matter more than a perfect checklist.
• A clear communicator. You document what you find, reproduce issues cleanly, and escalate with enough context for senior engineers to act quickly.
• Startup-ready. You thrive in ambiguous, fast-moving environments, wear multiple hats, and act like an owner when something is broken in the field or the lab.
• Mission-driven. You are excited about physical AI / robotics and want to help ship reliable hardware+software systems at scale.

Nice-to-haves:

Experience with camera systems, calibration (intrinsics/extrinsics, ArUco/ChArUco, OpenCV), ROS/ROS2, SLAM or computer vision projects, Jetson or similar SBCs, FreeRTOS / embedded Linux, PCB bring-up, soldering, Altium or similar EDA tools, competition robotics (e.g. ROBOCON), or prior work debugging consumer/industrial robots or wearables.

What You’ll Do:

• InstaCore Debugging. Diagnose and fix software and hardware issues across InstaCore devices — cameras, sensors, compute, firmware, cables, mounts, and power — from lab benches to deployed fleets.
• AI-Accelerated Engineering. Use Cursor, Claude Code, Codex, and similar tools as a force multiplier for scripting, log analysis, RCA write-ups, and rapid tool-building — while applying strong engineering judgment to validate results before they hit production hard…