IIIT NAGPUR · B.TECH ECE, MINOR IN IOT · 2022–2026

I decide what to build, and I build the AI systems that ship it.

Product management and generative AI engineering, from the same person. RICE and JTBD on one side of the desk; RAG pipelines, multi-agent orchestration, and LLMOps on the other.

Product pipeline AI engineering pipeline
Research Define Prioritize Design Deploy Evaluate Generate Retrieve PB Ship
01

Featured build — IntelliStack

A 4-phase production AI/ML platform, end to end: retrieval, agents, evaluation, deployment.

PHASE 1

Advanced RAG Engine

HyDE, Corrective RAG, Self-RAG, and hybrid BM25+dense retrieval fused with RRF. Parent-child chunking cut the "lost in the middle" problem.

LangChain · Qdrant · FastAPI
LangSmith · Docker
PHASE 2

Multi-Agent Research System

5-agent CrewAI system on a LangGraph state machine, with a critique-revision loop that auto-improves reports scoring below 7/10.

CrewAI · LangGraph · Mem0
FastAPI · SSE
PHASE 3

LLMOps Evaluation Pipeline

CI/CD gate that blocks deploys on RAGAS regressions. Prompt A/B testing via LangSmith Hub with auto-promotion of the winner.

MLflow · RAGAS · DeepEval
LangSmith Hub
PHASE 4

Cloud Deployment

Terraform-provisioned AWS EKS, zero-downtime rolling deploys, KEDA autoscaling to zero overnight, full Prometheus/Grafana observability.

Kubernetes · Terraform · AWS EKS
GitHub Actions
0.886
RAGAS score, hybrid+rerank vs 0.789 dense-only
+12%
faithfulness improvement over naive RAG
218s
end-to-end multi-agent research run
5,000+
lines of production Python, open-sourced
02

Product case studies

Independent teardowns and PRDs applying real PM frameworks to real products.

JTBDRICE

UPI Split Payment

Mapped drop-off points across GPay and PhonePe's group bill-splitting flows. Wrote a full PRD with a RICE-prioritized backlog and a 5-screen Figma redesign.

Analyzed pain points across 720M+ daily UPI transactions
AARRRMoSCoW

CRED — Product Teardown

Full-funnel teardown of CRED's rewards ecosystem, from onboarding to redemption. Identified two critical UX gaps and ranked 8 proposals by go-to-market impact.

Mapped the journey for 35M+ verified users
JTBDRICE

EdTech Engagement Drop-off

Investigated why course completion sits below the 15% industry average. Proposed a streak-based accountability system with peer cohorts and revision nudges.

3 personas · North Star: Weekly Active Learners
03

Two toolkits, one person

The product side and the engineering side, kept genuinely separate — not blended into vague "full-stack" language.

Product

RICE Prioritization Jobs-to-be-Done AARRR MoSCoW North Star Metrics OKRs Opportunity Solution Tree PRD Writing Wireframing Figma Notion Jira

AI Engineering

LangChain LangGraph CrewAI RAGAS / DeepEval MLflow Qdrant / FAISS / Pinecone FastAPI Docker / Kubernetes Terraform AWS EKS Prometheus / Grafana PyTorch
04

About

I'm a B.Tech student at IIIT Nagpur (Electronics & Communication, minor in IoT), working the product problem from both ends. On the product side, I run structured research, write PRDs, and prioritize with RICE and JTBD — mostly in FinTech and EdTech, where the constraints are real and the data is messy. On the engineering side, I build the actual AI systems: RAG pipelines that get evaluated against RAGAS before they ship, multi-agent systems with real critique-and-revise loops, and the Terraform/Kubernetes infrastructure to run it all in production — not just in a notebook.

2023 — Present

Core Member, Public Relations Lead — E-Summit

E-Cell, IIIT Nagpur

Led PR strategy for E-Summit — outreach, partnerships, and communications driving event awareness. Core contributor to campus entrepreneurship culture through workshops and speaker sessions.

2022 — 2026

B.Tech, Electronics & Communication Engineering

Minor in IoT — Indian Institute of Information Technology, Nagpur
Google Project Management Certificate2024
IBM Business Management Certificate2024
IBM AI & Product Management2024
Figma UI/UX Design2024–2025
Notion Essentials2024