• Built Roger’s core AI lead-finding engine from scratch, including intent-signal lead finding (hiring and competitor-follower signals), and scaled it while the company's paying B2B customers grew 20x. • Cut cost per lead by 10x with a multi-stage retrieval funnel: LLM search planning into firmographic filters, Typesense vector search, a Jev (TypeSafe AI) pre-filter, OpenSearch kNN people search, reranking, LLM fit/competitor/role filters on surviving candidates, and email enrichment. • Ran model evaluations on hand-labeled data to migrate the LLM filters to a 5x cheaper model with no drop in filtering accuracy, tuned vector and rerank thresholds for ∼95% recall on a huge candidate pool, and rebalanced the funnel to preserve precision. • Ran a 3-node OpenSearch cluster with 400M+ people records and built its monthly ingestion pipeline on Azure infrastructure. • Engineered the engine for production on BullMQ/Redis: resumable, idempotent paginated jobs with per-page checkpoints, per-job cost limits, auto-retirement of unproductive queries, queue prioritization, and graceful recovery. • Launched influencer marketing as a new outreach channel: customers find relevant LinkedIn creators (semantic search, LLM vetting, fake-engagement detection) and pitch them paid promos, with tracking of deliverables and outreach outcomes. • Built autonomous campaign management with AI agents: a daily LLM agent reviews each org’s campaign performance and edits or creates campaigns with validation and cooldowns, plus self-healing fallback campaigns when leads run out. • As lead engineer, reviewed other developers’ code across multiple services: the Next.js app, Express lead API and BullMQ workers.

Led a cross-functional team of 3 developers and 2 UX designers to architect and develop a Full-Stack Web App and cross-platform Mobile App with complex business logic. Conducting regular code reviews, enforcing coding standards, and ensuring high code quality aligned with project requirements. Managing end-to-end productionization, deployment, scalability, and release processes to successfully launch the application to market.

Developed full stack AI/LLM-based web apps using Next.js, TypeScript and Nest.js. Enhanced the user experience of key AI voice features on a leading Voice AI platform. Contributed to the development of an AI-powered collaborative rich-text editor with real-time features. Built an Asset Management System using AWS S3 and Next.js for handling uploaded and AI-generated assets.

Implemented end-to-end user authentication and verification using third-party APIs. Developed a feature-rich dashboard with complex logic for slot booking, service management, and scheduling. Optimized database operations using Prisma with PostgreSQL, achieving a 50% performance boost.