Work Experience

My work experiences across different companies and roles.

Peak Agent AI logo

Peak Agent AI

Full Stack Developer

Sept 2025 - Apr 2026

England, UK (Remote)

Technologies & Tools

Next.jsPythonTypeScriptSupabasePostgreSQL

What I've done

  • •Built a Stripe-integrated subscription billing system with webhook-driven plan synchronization, checkout deduplication, and a usage-based LLM credits metering system.
  • •Developed end-to-end authentication and onboarding flows using Supabase Auth with email OTP, Google OAuth, org provisioning, and reCAPTCHA v3 bot protection.
  • •Engineered an AI ad creation pipeline generating 3–4 concepts per request using GPT-5 for ideation and GPT-Image-1.5 with Replicate fallback for rendering.
  • •Built product page extraction service using Playwright with HTML pruning and LLM-based parsing to extract product features and images for ad generation.
  • •Architected a multi-agent orchestration system in Python with OpenAI function calling, enabling autonomous and scheduled decision-making over live ad platform data.
  • •Developed a Meta competitor radar platform tracking 100 ads via Ad Library scraping, with LLM-powered hook detection on competitor creatives.
  • •Designed a Redis-backed job scheduling system with user-reviewable job specs, ensuring controlled execution of agent-driven tasks across asynchronous workflows.
Infograins logo

Infograins

Full Stack Developer Intern

Jan - Apr 2025

India (Remote)

Technologies & Tools

Next.jsTypeScript

What I've done

  • •Set up a production-ready environment on AWS EC2 with Nginx reverse proxy and HTTPS using Let's Encrypt, ensuring 99.9% uptime.
  • •Reduced form submission errors by integrating React Hook Form with schema-based validation using Zod, providing real-time feedback and improved UX.
  • •Created a note-taking and document Q&A platform using Next.js and OpenAI APIs, enabling users to interact with PDF content contextually.
  • •Built a RAG-based PDF Q&A system leveraging vector embeddings and semantic search, allowing users to upload documents and ask context-aware questions with accurate responses.