Divya Prajapati

Ai Developer @Aurus Inc

Norwood, MA, US

Built a config-driven, parent–child LLM platform (Aurus) where shared core + adapters power multiple business cases (APDP, Ticketing) with zero code forks—new bots spin up by adding config, prompts, and secrets. • Engineered end-to-end RAG pipelines: standardized data collection/normalization to JSONL, schema validation, and reproducible ingestion flows for reliable knowledge grounding. • Productionized MongoDB (Atlas) for multi-tenant isolation: DB-per-case-per-env, least-privilege users, required indexes, and Atlas Vector Search/FAISS for scalable, low-latency retrieval. • Delivered a FastAPI service layer with /chat,/ingest,/health,/metrics, streaming responses, guardrails, and observability (tokens, latency, retrieval hit-rate). • Automated CI/CD in GitHub Actions: matrix runs across (case × env), nightly ingests, and per-environment secrets—drastically reducing time-to-launch and ensuring consistent, secure deployments.

Mumbai University Mumbai

Postgraduate Degree, Information Technology

Mumbai University Mumbai

Bachelor's degree, ]

Northeastern University

Master's degree, Computer Software Engineering