ms

✦ Hello, I'm

Mangalam Shrivastava

AI engineer & full‑stack developer

I build RAG pipelines, chat assistants with memory, and the guardrails that keep them safe.

Say hello

About me

I'm a computer science student at Vellore Institute of Technology who loves the place where AI meets real software. At Relevance Lab I helped ship a production RAG chatbot that reads thousands of documents and answers safely. On my own time I built Luna, a gentle health assistant inside a period‑tracking app, and a student portal that stays calm on registration day.

  • StudyingB.Tech CSE, VIT · Class of 2027
  • FocusLLM apps, RAG, safe AI
  • CurrentlyFinishing two certifications

Experience

Artificial Intelligence Intern

Relevance Lab Pvt. Ltd.

10,000+multi‑format documents ingested
40%higher model inference throughput
35%lower average API response latency
99.8%of prompt‑injection vectors mitigated
  • Engineered a production‑grade knowledge‑aware conversational agent (RAG chatbot) with a team of six engineering interns, optimising the data ingestion pipelines behind it.
  • Built FastAPI and Python backend services that orchestrate asynchronous LLM calls and multi‑turn chat memory, with thread‑safe data flow into a ChromaDB vector store.
  • Added context validation and security guardrails, and designed an agentic web retrieval layer with BeautifulSoup that scrapes, summarises and reflects on external sources.

Projects

GenAI · Health

Luna, a predictive menstrual health platform

A web period‑tracking ecosystem with cycle visualisation, predictive analytics and an in‑app assistant named Luna.

  • Designed Luna with LangChain and carefully engineered system prompts to give contextual health insights from each user's metrics.
  • Built a FastAPI backend for LLM integration, context‑window management and multi‑turn conversation state.
  • Added strict input validation and content‑safety guardrails across the LLM lifecycle.
PythonLangChainOpenAI APIFastAPIReactMongoDB
Full stack · Performance

High‑concurrency student portal

A fast replica of a university student portal, rebuilt to stay smooth under peak registration traffic.

  • Designed a full‑stack replica that removes the concurrency bottlenecks students hit during registration.
  • Built a scalable Node.js and MongoDB backend for high‑volume transactional data with lower API latency.
  • Developed a responsive React interface with optimised state management for smooth real‑time updates.
Node.jsExpressMongoDBReactPostman

Skills

Languages
PythonJavaScript (ES6+)HTML & CSS
AI & LLM engineering
RAG pipelinesPrompt & context engineeringAgentic reasoningGuardrails & injection defence
Frameworks
LangChainFastAPIStreamlitNode.jsExpress.jsReactBeautifulSoup
Data & vectors
MongoDBChromaDBSQLite
Tools & DevOps
Git & GitHubREST APIsDockerAWS EC2Postman

Certifications

In progress

Certification name

Issuing organisation · Expected 2026

In progress

Certification name

Issuing organisation · Expected 2026

Education

  • 2027

    Vellore Institute of Technology

    B.Tech in Computer Science and Engineering
    CGPA 8.78
  • 2023

    Kendriya Vidyalaya

    Senior Secondary (Class 12)
    86%

Let's make something lovely

I'm open to internships and full‑time roles in AI and full‑stack engineering. The quickest way to reach me is email.

LinkedIn GitHub Download résumé