Bengaluru, Karnataka, India
I build AI-first B2B products from zero to first revenue - and I've done it end-to-end, from writing the first POC to closing the first enterprise contract. In under two years, I went from intern to sole product owner of an entire B2B SaaS vertical. In that time I: → Closed my company's first SaaS deal (₹1.5 Cr ARR) after years of failed attempts on the same product → Led the AI product turnaround behind a Fortune 500 manufacturer's first-ever Global Channel Partnership in India - a 3-year contract at ₹40 Cr/yr → Cut a core operational workflow from 6-8 weeks to 9-14 days using multi-agent AI systems → Recovered a stalled enterprise engagement - process accuracy 78% to 94%, TAT 14 days to 9 hours I work fluently across the AI stack : multi-agent orchestration (LangGraph, Google ADK), hybrid RAG (Qdrant, Voyage, BM25), OCR + LLM pipelines, and production-grade evaluation. I prototype these systems myself and hand them to engineering. Code is on my GitHub. What I care about: turning AI capability into measurable business outcomes - revenue, accuracy, speed, scale. The domain doesn't matter to me; the problem-solving does. I've done it in B2B procurement and I'm ready to do it in any product environment. Currently looking for PM and AI PM roles at ambitious B2B/B2C SaaS companies.
Sole product owner of Venwiz's Pre-PO vertical post company-wide layoff. • Closed Venwiz's first SaaS deal — ₹1.5 Cr ARR with Cyient including a separate VendorIQ module; first SaaS revenue in the product's 4+ year history. • Recovered a stalled enterprise engagement with a Fortune 500 manufacturer — extraction accuracy 78% to 94%, processing TAT 14 days to 9 hours, engagement hit 130%+ of quarterly target; converted into a 3-year managed-service contract at ₹40 Cr/yr. • Aligned stakeholders for the RFQPro SaaS pivot despite competing company priorities, built the roadmap, and personally drove the client demos that landed Venwiz's first SaaS pilot. • Built production-ready multi-agent AI POCs handed to engineering with minimal rework — PQR generation (Google ADK + hybrid RAG on Qdrant + Voyage + BM25) and Technical Bid Analysis (LangGraph CRAG with critic-node corrective loop + Claude Vision OCR fallback). Code on GitHub. • Took full ownership of VendorIQ post-layoff — a GSTIN-based vendor risk and business-health product comparable to Probe42 / D&B, including backend, frontend, and data flow contracts. • Managed Venwiz's full ClearTax API engagement — 8 APIs powering RFQPro, VendorIQ, and Lead Generation — owning annual consumption forecasts and per-API-call pricing negotiations. • Initiated and ran weekly AI sessions across Product + Tech for 6+ months — accelerated TBA automation in Pre-PO and drove architecture changes in our enterprise chatbot and Post-PO manufacturing-plan automation.
Expanded from Search to full Pre-PO vertical ownership after PM attrition. • Identified vendor discovery as the bottleneck blocking a 90–100 RFQ/month client onboarding. Scoped and shipped an end-to-end agentic discovery + PQR automation system in under 2 weeks (POC to production). Scaled team capacity from 30–40 to 90+ RFQs/month and cut RFQ TAT from 6–8 weeks to 9–14 days. • Led Megalith — a vendor data mapping engine. Personally designed mapping logic (PAN to website, PAN to Indiamart, website to GSTN); expanded the vendor database from 19,000 to 2.5 lakh profiles in 6 months, selectively enriching Search. • Hired and managed 2 interns — one on Invoice Reading App development, one on vendor category search.
Search product ownership under senior PM. • Diagnosed Neo4j as an architectural mismatch for weighted multi-field vendor search; led migration to Elasticsearch — studied Apache Lucene internals, defined MongoDB fields to index, and tuned field weightages with engineering. Lifted Search's share of client-selected vendors from 40% to 75%. • Shipped a Retool-based internal search tool powered by the new ES algorithm, replacing fully manual vendor lookup across the operations team.
Built a personalized food recommendation system using embeddings (Chroma + OpenAI) on Recipe1M. • Developed a nutrition computation pipeline using USDA dataset, enabling accurate macro/micro analysis for recipes • Optimized large-scale (14M datapoints) nutrition processing using NLTK-based methods, reducing latency by 3.5×
Processed the data, released through RTI Act, into machine readable form for the India Data Portal, facilitating public access. • Utilized numpy and pandas python packages for data cleaning and manipulation to transform wide data to long data • Used process.extractOne module of fuzzywuzzy python library to extract the best matches of unmapped data and its score • Constructed a python script to map processed and unmapped data, scorewise, for each hierarchy of the Panchayat system
Students' Alumni Cell team works under the Dean of Alumni Affairs. The team strives to establish and nurture the interaction between the Institute and the alumni. It has served as a platform facilitating communication between the current students of IIT Kharagpur and their predecessors.We organize events like the Annual Alumni Meet (Reunion), Homecoming, Regional Student-Alumni Meet, Leadership Summit, Alumni Talks and carry on programmes like the Alumni Student Mentorship to meet our objective.