codeguardians
ford's multi-agent code-review platform.
projects
swe-term
a terminal swe-agent harness whose go core owns the agent loop while heavy engines run as sidecar binaries.
kairos
a contextual bandit for bookmark surfacing that learns when to interrupt, not only what matches.
obsidian agent
a langgraph research pipeline that turns a topic into a report, atomic notes, and durable memories through sequential agents.
antisomnus
a real-time driver-fatigue detection pipeline using face landmarks and cnn embeddings, deployed on gcp with a flutter client.
areas
- agentic systems
- code intelligence
- retrieval and context engineering
- llm evaluation and observability
- ai and cloud infrastructure
writings
curriculum vitae
AI Engineer shipping LLM workflows and agentic toolkits that augment software engineering. Focused on multi-agent orchestration, hybrid RAG, distributed inference, and developer-in-the-loop automation.
Experience
AI/ML Software Engineer
- Owned architecture and infrastructure for a LangGraph multi-agent pull-request review GitHub App, scaling to more than 1,000 product teams and 800–850 pull requests per day.
- Built diff-aware context and retrieval over repository and pull-request artifacts with multi-hop reasoning to detect actionable anti-patterns beyond CI linters; achieved 70% precision, 56.5% recall, and 59.4 F1 per pull request.
- Led platform hardening across Cloud Run services and tooling, human-in-the-loop escalation, and an in-house evaluation harness with guardrails and tracing through Arize Phoenix and OpenTelemetry.
Software Engineer
- Led an EV battery Track & Trace platform with a tamper-evident ledger to improve supply-chain transparency and auditability.
- Drove requirements with more than five internal and external stakeholders and translated them into specifications and UI workflows, contributing to a 25% increase in EV client satisfaction scores.
Software Engineer Intern
- Helped refactor a monolithic vehicle-diagnostics application into domain-aligned services using domain-driven design patterns and delivered API and frontend work in an Agile team.
Education
University of Michigan
Bachelor of Science in Computer Science & Data Science
Societies: Upsilon Pi Epsilon; Michigan Student Artificial Intelligence Lab.
Certifications
- AI Evals for Engineers & PMs
- Mastering LLMs for Developers & Data Scientists
Technical index
- PyTorch, TensorFlow, Hugging Face, LangChain, LangGraph, vLLM, Ray, Airflow
- FastAPI, Spring Boot, Docker, Kubernetes, OpenShift, Terraform, GCP, Azure
- SQL, Spark, BigQuery, Arize Phoenix, OpenTelemetry, Git
reading
Loading books…
Bookmarks
Loading bookmarks…
Hey, I'm Rohit.
AI Engineer
minds ∩ machines