FaiqueAhmadKhan
Full-stack engineer crafting fast, thoughtful products for the web — currently shipping AI experiences to 200k+ users.
About
Full-stack engineer with 3+ years building AI-integrated products used by 200K+ users, where clean architecture and thoughtful design meet real production scale.
Specializing in React, Next.js, and Node.js, with deep experience in LLM orchestration, including architecting a unified abstraction over 27+ models (GPT, Claude, Gemini, DeepSeek) with dynamic routing for cost, latency, and reliability.
Shipped high-impact features end-to-end, from an AI search engine that drove a 35% revenue lift to a streaming rich-text editor with deeply integrated Claude API completions, plus GIS platforms, SEO pipelines, and client-facing products across the stack.
Currently exploring Go and distributed systems to deepen backend and infrastructure skills alongside my AI/LLM work.
Outside of work, I enjoy cycling, football, historical mystery films, and building AI-driven side projects.
Experience
- Built AI Docs from the ground up, a production-grade rich-text editor with deeply integrated Claude API, supporting streaming inline completions, context-aware content transformation, and multi-block editing.
- Designed the document state model to handle concurrent AI edits without corrupting user content.
- Shipped an AI Search Engine with real-time web retrieval, semantic query handling, and fallback ranking logic. It became the product's top conversion driver, contributing to a 35% lift in subscription revenue within 3 months of launch.
- Architected the LLM orchestration layer for Chatly AI (200K+ users), designing a unified abstraction over 27+ models including GPT, Claude, Gemini, and DeepSeek.
- Built dynamic model routing based on cost, latency, and capability tiers, with fallback logic that reduced failed completions by 20% and kept p95 response time under 8 seconds.
- Led a 3-person engineering team (including 2 junior engineers) on a GIS platform from architecture to production delivery.
- Introduced PR review standards and a branching workflow that cut integration conflicts by roughly half, allowing the team to ship on a fixed client deadline with no scope cuts.
- Architected a full-stack GIS platform from schema design to production, using Next.js, Node.js, and PostgreSQL for geospatial queries, with Firebase for real-time sync.
- Designed the event ingestion pipeline to handle 10K+ concurrent location events with sub-second UI updates.
- Optimized PostgreSQL spatial indexes to keep query latency under 100ms at load.
- Built and maintained CI/CD pipelines across Vercel and AWS, cutting deployment friction and ensuring zero-downtime releases for production environments.
- Owned end-to-end frontend delivery across 12+ client projects at a software agency, driving architecture decisions, component design, and production deployments independently with minimal oversight.
- Introduced TypeScript strictly and established reusable component patterns using Redux and Zustand, adopted across the team, reducing redundant code.
- Cut estimated feature delivery time on repeat project types by ~30% through reusable patterns and standards.
- Consistently delivered pixel-perfect Figma-to-production implementations across diverse client stacks (Tailwind, Material UI, Radix, Ant Design).
- Became the team's go-to engineer for high-fidelity UI work on tight deadlines.
Selected Work
Chatly AI
Unified platform combining AI chat, search, and image generation to streamline writing, idea generation, and document support.
Best Western Premier Islamabad
Luxury hotel website showcasing premium rooms, amenities, dining, wellness & event spaces, set against the backdrop of Margalla Hills.
ChatCoach.io
AI-powered virtual mentor that simulates real conversations to help users improve communication skills with emotional analytics and scenario-based coaching.
beFesti.nl
Dance music & festival culture platform oriented toward Gen Z; helps brands engage through content, exposure, and culture-driven campaigns.
Get in Touch
Elsewhere
Availability
I usually reply within 24 hours. For anything urgent, LinkedIn is the fastest way to reach me.