Google Robotics Jobs (NOW HIRING)

Location: Bay Area / San Mateo, CA

Employment Type: Full-Time

Department: Engineering

Deccan AI is a model training and evaluation startup headquartered in the Bay Area, with a delivery office in Hyderabad. We are founded by IIT Bombay, IIM Ahmedabad, and ex-Google alumni, and we work with leading AI labs and enterprise teams on high-quality human data, evaluations, and AI-first scaled operations.

We believe frontier AI systems are only as good as the data, evaluations, and human judgment behind them. Our work sits at the intersection of customer needs, research ambition, and operational execution. That makes the work high-touch, technical, ambiguous, and execution-heavy.

We are looking for people who can operate with urgency, write clearly, build trust, and turn ambiguity into outcomes.

We are hiring a Founding Engineer - Robotics to help build Deccan’s robotics and physical AI practice.

Robotics data is entering the phase where the right data infrastructure, quality systems, and workflow design matter as much as raw collection volume. Customers are asking for egocentric data, annotation, chunking, video / image work, teleoperation, industrial and retail settings, and defensible quality control. Deccan already has real robotics operations and customer demand, but the next phase requires stronger technical systems: better data specs, better collection and validation workflows, better tooling, and a clearer bridge between robotics customers, field operations, and model-development needs.

This is a hands‑on founding engineering role. You will build the technical foundation for robotics data and physical AI workflows: egocentric capture, multi‑modal data pipelines, task design, metadata standards, automated QC, VLM-based validation, annotation tooling, and systems that make robotics data usable for model training and evaluation. You should be comfortable working across software, computer vision, ML, data operations, and customer‑facing technical scoping.

• Build systems for robotics and physical AI data workflows: egocentric video, multi‑camera capture, annotation, chunking, metadata, task definitions, and quality checks.

• Translate customer asks into clear technical specs: collection environment, task inventory, sensor requirements, output format, labeling taxonomy, QC criteria, and production workflow.

• Design tooling that helps Deccan scale beyond household tasks into retail, industrial, manufacturing, warehouse, and teleoperation settings.

• Build automated QC systems for robotics data, including heuristic checks, deduplication / fraud detection, illumination and cut detection, metadata validation, and VLM-based validation against task descriptions.

• Partner with field operations and delivery teams so collection workflows are technically sound, repeatable, and measurable.

• Partner with GTM and Engagement Managers to make Deccan’s robotics capability credible in customer conversations without overpromising.

• Prototype quickly, then harden the components that need to become reusable infrastructure.

• Create playbooks for robotics data projects: scoping, task design, capture setup, contributor instructions, QA / QC, acceptance criteria, and delivery handoff.

• Help decide where Deccan should build tooling, partner with hardware providers, or standardize capture modalities.

• Stay close to the market in physical AI, world models, video understanding, robot learning, teleoperation, egocentric data, and multimodal model training.

We are looking for a strong engineer who can combine robotics intuition, computer vision / ML judgment, and systems‑building ability.

You should be:

• technical to design data systems for robotics, video, multimodal, and physical AI workflows;

• practical to turn messy customer needs into executable specs and tooling;

• quality‑obsessed to care about collection setup, task ambiguity, metadata, failure modes, and QC edge cases;

• systems‑minded to build reusable infrastructure rather than one‑off scripts;

• comfortable working with operations teams without becoming an operations manager;

• high‑agency to build the first version before all requirements are perfectly defined;

• clear in writing to create specs, instructions, playbooks, and customer‑facing technical explanations;

• direct to push back when a robotics ask is underspecified, not feasible, or unlikely to produce useful data.

• 4+ years of experience building software, ML, computer vision, robotics, data infrastructure, autonomy, simulation, or multimodal systems; exceptional earlier‑career candidates with strong evidence of ability will also be considered.

• Hands‑on experience with computer vision, video data, robotics data, sensor data, teleoperation, autonomy, physical AI, or world‑model workflows.

• Strong programming ability in Python and modern data / ML tooling; experience with PyTorch, OpenCV, VLMs, video processing, or data pipelines is a plus.

• Experience building data quality systems, annotation tools, evaluation harnesses, validation pipelines, or production data workflows.

• Ability to reason about task design, metadata, collection protocols, human‑in‑the‑loop quality, and model‑facing data requirements.

• Comfort working across engineering, delivery, operations, GTM, and customer stakeholders.

• Startup or zero‑to‑one experience where the job was to create structure, not wait for it.

• Bay Area presence or willingness to work closely with the San Mateo team.

At Deccan AI, you will work close to the frontier of AI model training, evaluation, and scaled human‑data operations. Robotics and physical AI are still early: data formats, capture methods, quality standards, and workflows are being defined now.

This is a high‑agency role for someone who wants to build those systems, not just discuss them. You will help define Deccan’s robotics practice, create reusable infrastructure, and work on problems where the boundary between software, data, operations, and physical‑world complexity is still being drawn.

If you are excited by robotics data, computer vision, physical AI, technical ambiguity, and the chance to build a practice from zero to one, we would like to meet you