I'm a software engineer who builds AI products and takes them all the way to production: from the first call with a customer, through system design, APIs and databases, to deployment, monitoring and Day-2 support.
My day-to-day is solid backend engineering in Python and FastAPI, React frontends and PostgreSQL on AWS, combined with applied AI: LLM agents, RAG and real-time voice where the product needs it. I've shipped production systems for remote teams in 🇺🇸 the US, 🇦🇺 Australia and 🇸🇪 Sweden.
- 🔭 Currently AI Engineer at TarjamAI, owning a B2B real-time translation platform end to end
- 🧱 Comfortable across the stack: async Python services, REST APIs and webhooks, React dashboards, SQL schema design, CI/CD
- 🤝 Customer-facing by habit: I turn fuzzy requirements into technical designs, integrations and shipped features
- 🎓 Pursuing an MS in Artificial Intelligence at NED University
- 🧑🏫 Mentored 200+ students across 3,000+ Python & data science sessions (4.9★)
- 💼 Open to Software Engineer, AI Engineer and Forward Deployed Engineer roles
| 🧱 Software Engineering | 🤖 Applied AI & Agents | 🚀 Forward Deployed |
|---|---|---|
| Full-stack products on FastAPI, React and PostgreSQL. API and schema design, async services, testing, CI/CD and production monitoring. | LLM agents with tool calling, RAG, multi-agent orchestration, MCP, and real-time voice AI on LiveKit and Twilio SIP. | Working directly with customers: discovery, requirements, third-party API integrations, deployment and Day-2 support. |
| 🟢 99.5% | 👥 3,000+ | 🐞 −35% | ⚡ < 800 ms |
|---|---|---|---|
| uptime on production Python APIs | users on AI products I shipped | production error rate after fixing 40+ critical bugs | end-to-end latency on real-time voice pipelines |
🇺🇸 United States (Remote) · Apr 2026 – Present
B2B platform sold to masjids that translates live Friday khutbahs and Islamic sermons into 100+ languages in real time.
- Own the product end to end: discovery, system architecture, implementation, deployment and Day-2 operations
- Architected a low-latency streaming speech-to-translation pipeline that keeps live captions in sync with the speaker through long-form sermons
- Grounded output in authentic sources by detecting Quranic verses, hadith and Islamic terms mid-speech and linking them to Quran.com and Sunnah.com
- Work directly with masjid administrators to turn their needs into requirements, onboarding flows and a prioritised roadmap
System Design Streaming Pipelines Deployment & Monitoring Customer Discovery
🇦🇺 Australia (Remote) · Nov 2025 – Apr 2026
- Built and maintained Python REST APIs and automation workflows, sustaining 99.5% uptime across production systems
- Integrated the Mews hotel reservation API as agent tools for creating, updating and cancelling bookings and fetching guest profiles, removing ~15 hours/week of manual work for hospitality clients
- Diagnosed and fixed 40+ critical production bugs in LiveKit voice agents, cutting error rates by 35%
- Introduced automated pre-release testing with HammingAI, raising call-completion accuracy by 22%
Python REST APIs Third-party API Integration LiveKit Agents Automated Testing
🇸🇪 Sweden (Remote) · Jul 2025 – Nov 2025
- Shipped multilingual healthcare chatbots with contextual memory, serving 3,000+ users in 5 languages
- Built a multi-agent system for question answering over SQL databases, cutting data-retrieval time by 45%
- Integrated AI ticket triage into the support workflow, deflecting 60% of tickets away from human agents
LLM Agents Multi-Agent Systems SQL Tool Calling Workflow Integration
🇵🇰 Karachi, Pakistan · May 2025 – Oct 2025
- Integrated Twilio SIP with LiveKit for inbound and outbound calling, scaling reliably to 100+ concurrent calls
- Designed a streaming STT → LLM → TTS pipeline with real-time interruption handling at < 800 ms latency
- Implemented real-time event streaming and state synchronisation across services using data channels and async RPC
LiveKit Twilio SIP WebRTC Event Streaming Async RPC
🇵🇰 Karachi, Pakistan · Jan 2025 – Apr 2025
- Built generative-AI support workflows with LangChain and RAG, cutting average resolution time by 30%
LangChain RAG LLM Workflows
🌐 Remote · Mar 2022 – Present
- Mentored 200+ students through 3,000+ one-to-one Python and data science sessions, including learners from Harvard, Cornell and Johns Hopkins, reviewing their code along the way (4.9★)
Python Data Science Machine Learning Code Review Mentoring
flowchart LR
A["🔍 Discover<br/>requirements"] --> B["📐 Design<br/>APIs · schema"]
B --> C["🛠️ Build<br/>FastAPI · React · agents"]
C --> D["🧪 Test<br/>Pytest · CI"]
D --> E["🚀 Ship<br/>Docker · AWS"]
E --> F["📈 Operate<br/>Sentry · CloudWatch"]
F -.->|feedback| A
🧩 Side projects & freelance work
- MajsterAI: full-stack AI receptionist for trade businesses. FastAPI + PostgreSQL backend, React 18 + Vite dashboard, LiveKit voice agent, shipped as a pnpm monorepo with Docker Compose.
- Tixse: freelance event-ticketing web app with real-time seat selection that prevents double-booking, payment integration and AI-powered support.
| 🎓 MS, Artificial Intelligence | NED University of Engineering & Technology · 2026 – Present |
| 🎓 BS, Computer Science (AI & Big Data) | University of Karachi · 2021 – 2025 |
| 🏅 Certified AI Developer | PIAIC (Presidential Initiative for AI & Computing) · 2020 – 2022 |
| 📜 Building Agentic RAG with LlamaIndex | DeepLearning.AI |
| 📜 Building Autonomous Agents | DeepLearning.AI |
| 📜 LangChain for LLM Application Development | DeepLearning.AI |
| 📜 Machine Learning | Kaggle / DataCamp |
Need someone to take an AI feature from idea to production, plug it into your existing systems, and keep it reliable once real users arrive? I'd love to hear about it.
📫 hammadfaheem5@gmail.com · 💼 LinkedIn
"AI systems should feel seamless, not scripted."



