Gurugram, Haryana, India
I build production agentic AI systems end-to-end. Currently AI Automation Engineer at Ken Research, where I architect and operate a multi-platform content distribution platform spanning 13 blog platforms (Medium, Substack, Dev.to, LinkedIn Pulse, WordPress, Blogger, Google Sites, HackMD, and more) and 3 social networks, managing 150+ accounts via persistent Playwright sessions. The system publishes 600+ blogs per day through a multi-agent pipeline (Search → Writer → Image → QA → Publishing) using Claude, OpenAI, and OpenRouter with key rotation across 15 providers. The platform has evolved from a TypeScript cron script into a full distributed system: Next.js dashboard for job submission and tracking, Django + Celery backend deployed on Render for async job queuing, and a distributed Playwright worker fleet running across team machines - since real browser automation can't run on cloud servers. Team members submit jobs via the dashboard, Celery queues them, workers pick them up, scripts execute, results sync back to the dashboard and Google Sheets. Other systems I've built at Ken Research: → Client Partner lead outreach - runs every 3 hours, validates emails, sends GPT-4o personalized outbound with a 3-step follow-up sequence → LLM brand visibility tracker - pings 25 different LLMs on schedule, captures via Helicone → SERP ranking and indexing tracker - monitors Google positions across published reports → Video automation pipeline - Grok-generated clips stitched into 25–30s shorts for autonomous posting Stack: TypeScript, Python, Django, Celery, Next.js, Playwright | Claude/OpenAI/OpenRouter, Helicone | n8n, Google Sheets API, MySQL, PostgreSQL | Linux, Git, Render | AI tooling: Claude Code, Cursor Open to conversations about agentic AI, production agent reliability, distributed browser automation, and roles where I can build similar systems at larger scale. Happy to swap notes with other folks shipping agents.