N Divij
AI Engineer · Document Intelligence · Evidence Traceability · Data Validation · Agentic Workflows
New Delhi, India · UTC+5:30ndivij2004@gmail.comgithub.com/N-45divlinkedin.com/in/divij45divijworks.com
Open to AI engineering roles · available now
About
AI engineer with two years in industry since Nov 2024, all of it remote for teams in San Francisco, Stockholm, Hyderabad and Delaware, plus 18 hackathon wins. I build document and data pipelines where every AI output traces back to its source evidence, with validation and quality checks built in. I learn new domains fast: healthcare compliance, legal tech and geospatial data.
Experience
AI Engineer · Kivor
Nov 2025 — 2026San Francisco, remote from New Delhi
- Linked each diagnosis to its supporting evidence with vector embeddings, giving the client auditable explanations.
- Built a research-backed anti-gaming evaluation system for healthcare compliance workflows, supporting AI governance.
- Turned a healthcare client's compliance workflows into agent designs, and led delivery as their main point of contact.
- Built a voice-based agentic outreach system and improved the STT/TTS pipelines behind a client's interview-training voice agents.
- Designed Redis and MinIO pipelines that fetch, structure and serve agent outputs from enterprise data sources.
- Worked on an enterprise agent-builder platform (workflows, orchestration, tool use) and mentored interns and junior developers.
Full-Stack AI Agents Developer · M0
Aug 2025 — Nov 2025Hyderabad, remote from New Delhi
- Built geospatial AI agent workflows for urban planning and real-estate intelligence: Nest.js backends with streaming and caching, and Mastra/Agno orchestration for reliable tool use.
Software Developer Intern → Consultant (full-time) · LawyalTech
Feb 2025 — Jul 2025Stockholm, remote from New Delhi
- Built a Botpress and Gemini document-automation chatbot for contract understanding, plus React/Firebase legal-workflow components.
Blockchain & Full-Stack Intern · SkyTrade
Nov 2024 — Jan 2025Delaware, USA, remote from New Delhi
- Researched location data for a blockchain startup; made large map files remotely queryable with DuckDB and mapped them in QGIS and Kepler.gl.
Projects
Parallax | Document AI, PDF extraction, output validation
Reads a PDF five independent ways and treats disagreements between readings as evidence, catching hidden text before it reaches a model; benchmarked on 12 AI models.
Pareis | Python, DataHub, dbt, MLflow, scikit-learn, PostgreSQL, Docker
Data-quality and lineage pipeline over 1.07M real transactions: dbt tests validate datasets, lineage-graph traversal traces failures to affected MLflow models (AUC 0.78 / 0.82), and a CI gate blocks retraining on bad data.
EverydayElastic | Python, FastAPI, Elasticsearch, Vertex AI, Gemini, Cloud Run
RAG copilot over 7,800+ technical documents with hybrid search, reranking, and grounded citations back to source; 1st place, Google Cloud AI Accelerate.
ElitOrc-AI | Next.js, Mastra, TiDB Vector Search, Groq, Cohere
Multimodal clinical assistant chaining medical image analysis, vector case retrieval, and LLM reasoning with persistent memory; Winner, TiDB AgentX.
Hackathons
18 hackathon wins, each linked to its public proof at divijworks.com. 1st place or winner: Google Cloud AI Accelerate (Elastic Challenge), Solana Radar (CUDIS track), TiDB AgentX (Best Open Source), Backblaze Generative Media (Feedback Prize), Build an Agent Hackathon at SF Tech Week, Solar mini Hacker House Singapore, Solana Privacy Hack (DeFi track), Adaption AutoScientist (Part 1). 2nd place or runner-up: ETHOnline, Helius, Doodles Dreamnet, BAND Dark Factory, MNEE. 3rd place: AMD DevMaster, HashKey Chain Horizon, Flare × Google Verifiable AI. Also a community favourite (Agents of the Permaweb) and Rasa Agent Challenge Top 10.
Technical skills
- ML / Evaluation:
- scikit-learn, XGBoost, TensorFlow, MLflow, pandas, NumPy, embeddings, hybrid search and reranking, eval design, benchmarking, AI governance
- LLM / Agents:
- Agentic workflows, RAG, tool calling, structured outputs, prompt engineering, MCP, WebMCP, LangChain, LangGraph, LlamaIndex, Mastra, Agno
- Data / Observability:
- Grafana, Prometheus, Loki, DataHub, dbt, OpenMetadata, data lineage and quality checks, PostgreSQL, Redis, MinIO, Elasticsearch, Milvus, TiDB, MongoDB
- Languages:
- Python, SQL, TypeScript, JavaScript, Rust, Go, Java, C/C++, R (coursework)
- Infra / Apps:
- Docker, Kubernetes, GCP (Cloud Run, Vertex AI), AWS Bedrock, Cloudflare Workers, Git, CI/CD, FastAPI, Nest.js, Next.js, React
Education
GB Pant Government Engineering College, New Delhi
B.Tech, Computer Science and Engineering · CGPA 8.71