Ahmad Raza

Co-Founder

pseudobytes • Temps plein

Oct 2024 - Present • 1 yr 7 mos

At Pseudobytes, we build production-grade AI systems, SaaS platforms, and automation tools for global clients across healthcare, real estate, legal tech, e-commerce, and recruitment. As Co-Founder, I lead system architecture, AI integrations, API design, and end-to-end product delivery — from first commit to live deployment. Key projects I've shipped: → Sonoria AI — Real-time Voice AI SaaS handling inbound calls for 100+ businesses. Built on OpenAI Realtime API, Twilio, and Django Channels. Reduced AI inference costs by 75% via MCP and context caching. Secured $50K in seed funding. → ParcelIQ — Real estate intelligence platform analyzing 100,000+ properties/month. Automated 7 GIS risk checks, GPT-4o Vision satellite analysis, and skip tracing. Turned 30-minute manual lookups into 2-second automated reports. Cut AI costs by 90%. → KodeReach — LinkedIn prospecting SaaS with 150+ paying clients processing 7,000+ prospects/month using Selenium, FastAPI, and LangChain-powered personalization. → ShamelaGPT — Arabic RAG platform processing 60M+ pages, reducing researcher time by 60%. → MomentPik — AI photo platform with real-time facial analysis, 40% performance improvement, and 99.9% uptime. Stack: Python · FastAPI · Django · LangChain · OpenAI · GPT-4o · AWS · Pinecone · PostgreSQL · React · Next.js · Docker · Twilio · Whisper Maintained 99.9% uptime across production systems handling 20,000+ daily API requests.

Full Stack Developer & AI Engineer

Techpal • Temps plein

Mar 2022 - Dec 2025 • 3 yrs 9 mos

At TechPal, a Bahrain-based software company, I led AI platform development delivering production systems across thesis management, facial analysis, AI content generation, and photo intelligence. As Full Stack Developer & AI Engineer, I owned end-to-end development — from backend architecture and AI integration to frontend delivery and AWS deployment. Key projects I delivered: → MomentPik — AI-powered event photography platform with real-time facial analysis built on PyTorch and OpenCV. Automated sorting of thousands of event photos in minutes, enabled searchable galleries, and supported photographer monetization through automated photo sales. Achieved 40% performance improvement and 99.9% uptime across 10+ clients. → AI Thesis Management System — Automated review, scoring, and evaluation workflows for academic institutions, replacing manual processes with intelligent pipelines. → AI Course Generation Pipeline — Automated end-to-end content production workflows, reducing content timelines by 80%. Full-Stack Architecture: Built and deployed platforms using FastAPI and Django on the backend, React and Next.js on the frontend, with Redis and PostgreSQL for optimized data pipelines — achieving 40% improvement in overall system performance. All systems deployed on AWS with Docker, maintaining high availability and production-grade reliability across multiple concurrent clients. Stack: Python · FastAPI · Django · React · Next.js · PyTorch · OpenCV · AWS · PostgreSQL · Redis · Docker