Ahmad F.

As co-founder and CTO of Mergestack, I've architected and shipped 40+ production projects, including SaaS platforms scaled to 100k+ monthly users, MVPs backed by Google and MasterCard, and AI-powered tools used by 500,000+ practitioners. I work end-to-end: architecture, frontend, backend, AI integration, and deployment.

If your product needs to scale, perform, and ship, that's where I come in.

𝐖𝐇𝐀𝐓 𝐈 𝐁𝐔𝐈𝐋𝐃 𝐀𝐍𝐃 𝐒𝐇𝐈𝐏:

→ Full-Stack Web Applications
React, Next.js, Node.js, NestJS, Python, FastAPI, TypeScript; SaaS products scaled from 0 to 100k+ monthly users

→ SaaS Architecture and Multi-Tenant Platforms
Role-based access, isolated environments, subscription billing, and data models built for growth

→ AI Integration, Agents, and RAG Systems
OpenAI, Claude, LangChain, Pinecone, vector search; production-grade AI features, not demos

→ Pixel-Perfect Frontend Development
Figma to code, responsive, WCAG-compliant; 95+ Lighthouse scores, sub-1-second load times

→ Geospatial and Data-Heavy Dashboards
Mapbox, DeckGL, Google Maps; tested on 300k+ feature datasets with real-time interaction

→ RESTful and GraphQL APIs
150+ endpoints built with caching, rate-limiting, and webhook support

→ Third-Party Integrations
Stripe, Twilio, OpenAI, Google Calendar, Office 365, Shopify, Slack, DocuSign, SendGrid, and more

→ MVP Delivery
20+ MVPs launched in 6-12 weeks from scratch to production

𝐑𝐄𝐂𝐄𝐍𝐓 𝐖𝐎𝐑𝐊:

𝟏. 𝐀𝐈 𝐂𝐄𝐎 (𝐀𝐈-𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐑𝐀𝐆-𝐁𝐚𝐬𝐞𝐝 𝐂𝐨-𝐩𝐢𝐥𝐨𝐭 𝐟𝐨𝐫 𝐄𝐱𝐞𝐜𝐮𝐭𝐢𝐯𝐞𝐬):
Built an agentic AI system that acts as a conversational co-pilot for company leadership, giving CEOs direct, natural-language access to company data, reports, and operational context without waiting on manual reporting cycles.
React, LangChain, RAG, Pinecone, OpenAI/GPT-4, Agentic workflow orchestration

Designed a multi-step agent architecture (not single-shot Q&A) capable of reasoning across multiple internal data sources - financials, operations, and business metrics - to answer complex executive-level questions in plain language. Built a retrieval layer over structured and unstructured company data so the agent could ground its answers in real, current information rather than generic model knowledge. Focused heavily on making the agent trustworthy for high-stakes decision-making: clear source grounding, guardrails against hallucinated answers, and an interface built for non-technical executive users. The client has continued expanding the platform's scope post-launch, adding new data sources and agent capabilities as adoption has grown internally.