About

Hi, I'm Bhargav.

Full-Stack Developer based in Berlin. Before the Master's, I spent years inside Fortune 500 infrastructure in Bangalore, performance monitoring, legacy rewrites, the kind of codebase where a slow query at 2am becomes someone's incident report by 9.

I moved to Germany and started building for early-stage products. The problems got smaller and sharper. For a VC fund I built a startup discovery engine that pulls from CrunchBase, LinkedIn, job boards, and a handful of scraped sources, deduplicates across them, and surfaces signals a human analyst would take a week to find. I've deployed RAG chatbots on vector databases where retrieval quality is the product, not a feature. Performance optimization at this scale means obsessing over p95 latency, not just averages.

At acemate.ai, an AI-powered study platform, I work across the stack and directly with the CXO on product, growth, and GTM. That means fixing SEO architecture, cutting client feedback cycles from days to seconds with Featurebase, wiring PostHog and Attio together for a clean customer data pipeline, and building email automation on Listmonk with Supabase Edge Functions. I also write for the blog when the marketing team needs someone who can explain technical things without making them sound technical.

Alongside this, my thesis. I'm building the Prospect Certainty Framework to detect LLM hallucinations using logit masking, reading the model's probability distribution as it generates tokens and flagging low-confidence spans before they reach the user. Existing evals catch hallucinations after the fact. This catches them during generation. Slow research, lots of dead ends, exactly the kind of problem worth sitting with.

I shoot on film when I can. Running clears my head faster than anything else.