Rakesh Singh

Resume

I'm Rakesh Singh, an AI Backend Engineer at Genpact based in Bangalore, India. I architect backend systems and performance-driven applications designed for scale and production stability — most recently grounded document Q&A systems that answer only with faithful citations and abstain when the evidence isn't there, and before that, Java and Spring Boot services processing millions of transactions a day for Shutterfly in the US.

My focus areas are scalable backend systems, production reliability, and — currently — retrieval-augmented generation done in a way that doesn't quietly lie to users. I'm open to selective remote opportunities.

Experience

AI Backend Engineer · Genpact

Bangalore, India

  • Designed idempotent PDF ingest with content hashing and stable chunk IDs so a changed document re-indexes without duplicate vectors or stale citations
  • Built authorization-aware hybrid retrieval (pgvector + BM25 + RRF, then cross-encoder rerank) scoped by user and workspace from day one
  • Shipped structured answer-with-citations and calibrated abstention as the default path, held on a frozen 40-question set (18 answerable / 11 partial / 11 unanswerable)
  • Modeled threats with DFD + STRIDE and ran a separate adversarial suite covering planted instructions, prompt extraction, cross-user access, and hostile PDFs
  • Added a bounded LangGraph path only for question types hybrid RAG still failed, with loop limits and a single read-only MCP document-lookup tool
  • Built eval gates separating retrieval quality, citation faithfulness, abstention precision/recall, security pass rate, and cost per successful task, with version tags

FastAPI · pgvector · LangGraph · evals

Backend Java Developer · Genpact (Shutterfly, USA)

Bangalore, India

  • Architected Java / Spring Boot microservices processing 2 million daily transactions, reducing API response times by 65%
  • Implemented Resilience4j across 15+ microservices and reached 99.9% system uptime
  • Optimized PostgreSQL, MongoDB, and Redis, increasing retrieval speed by 70% and cutting costs by 30%
  • Built a photo pipeline for 5 million daily uploads with sub-200ms responses and compression that cut storage cost by 40%
  • Mentored 3 developers and raised code coverage from 65% to 92% through TDD

Java · Spring Boot · PostgreSQL · Redis

Education

Bachelor of Technology, Computer Science and Engineering

National Institute of Technology (NIT) Bhopal

  • Foundation in OS, networks, databases, and software engineering.
  • Helped organize ISTE chapter events for 200+ students, and contributed 100+ volunteer hours at Arushi NGO, Bhopal.

Skills

Backend Systems

Production-grade distributed systems with sub-100ms response targets

FastAPI · Python · Java · Spring Boot · TypeScript

Retrieval & AI

Hybrid retrieval, citation validation, and bounded agent workflows

pgvector · BM25 · LangGraph · MCP · Structured Outputs

Languages

Core engineering logic in high-level and system languages

Python · Java 11/17 · TypeScript · JavaScript · SQL

Data & Storage

Schema design and transactional synchronization at scale

PostgreSQL · pgvector · Redis · MongoDB · SQL

DevOps & Quality

CI/CD pipelines with Docker, Kubernetes, and AWS infrastructure

Docker · Kubernetes · AWS · Jenkins · Resilience4j