AI Engineer

At O₂ AI, we are building a Vertical AI Agent for the global semiconductor and electronics supply chain, just as Harvey is transforming the legal industry.

Company Website: (Use the "Apply for this Job" box below)./

We build deeply vertical AI Agents capable of handling high-stakes, high-complexity challenges in global manufacturing systems, transforming chaos into efficient, intelligent automated execution.

We are looking for a hands-on Builder to own product engineering end-to-end: from understanding real customer problems, to designing and shipping AI-native systems, to iterating in production.

Work alongside founders & early members from Yale, Northwestern, NUS, Columbia, Stanford and MIT, with experience across Deep Seek, NVIDIA, and Microsoft, etc.. We are a team that cares equally about frontier AI and real product execution.

We ship: (1) systems used by real customers, (2) awarded publications at top venues.

At O₂ AI, we provide you with direct opportunities to communicate with customers. You will interface directly with US clients and collaborate with Fortune 500 electronics and semiconductor companies. Our products span three layers:

• Electronic Vertical AI Copilot — domain-specific knowledge assistance that understands industry context deeply enough to be trusted by experts.
• ERP Workflow Agent — automating complex operational workflows with real permissions, real data flows, and real state management.
• Pricing Decision Agent — high-stakes decision support where accuracy and reliability directly affect business outcomes.

Your code will touch real users, real operational constraints, and real revenue.

At O₂ AI, we create opportunities for you to deeply engage in product and engineering. You will participate in the complete process from problem definition and solution design to product launch and continuous iteration, gaining highly hands-on practical experience and witnessing the real-world impact of your work in customer business scenarios.

O₂ Lab, our downstream research-driven lab, collaborates deeply with Northwestern MLL Lab and MIT Media Lab. Members of the founding team have been recognized at venues such as NeurIPS LAW (Outstanding Paper) and ICCV SP4V (Best Paper), with 10k github stars, Huggingface #2 paper of the day, and invited talks at Deep Mind, MIT, etc.

We focus on translating research into production-grade vertical AI systems. You will have opportunities for joint research, publications, and turning research insights into shipped products.

Participate in end-to-end product and engineering workflows. You are expected to participate in at least one of the following terms in the product lifecycle:

• Agent architectures: context engineering, harness engineering (orchestration logic, routing, fallback chains, state management across turns).
• Backend Services: implement tool/function calling integrations (APIs, databases, internal services) and handle the messy edge cases: timeouts, malformed responses, cost tracking.
• Evaluation pipelines : think about failure modes like prompt injection, data leakage through context, runaway agent loops, cost blowups.
• Research-driven exploration : LLM agent training, reinforcement learning, continual learning for evolving domains, improving performance and publishing papers / open-source projects.

• Design and implement REST/GraphQL APIs with proper auth (OAuth2/JWT), input validation, and rate limiting.
• Write and migrate SQL schemas, normalize where it matters, denormalize where performance demands it.
• Implement RBAC or ABAC permission models that actually map to business requirements.

• Containerize services (Docker), write CI/CD pipelines, deploy to AWS/GCP.
• Set up logging, metrics, and alerts.

Strong coding fundamentals: at minimum, solid in Python. You can read, write, debug, and reason about code under pressure, with strong experiences coding agent workflow, training, or evaluation.

AI-native building…