Full Stack AI Engineer · Founder @ ZorceX · AI Automation & Agent Systems
From datasets to inference to user-facing product — AI agents, RAG pipelines & full-stack products for founders, startups & global clients.
I'm a Full Stack AI Engineer and founder of ZorceX, focused on taking machine-learning systems all the way to production — RAG pipelines, LLM agents, AI automation, and full-stack delivery. I've shipped work as an AI Engineer on Upwork and as an AI Agent & Automation Specialist on Fiverr, serving international clients, founders, and startups across NLP, workflow automation, and production AI APIs.
- 🎓 Pursuing a BS in Artificial Intelligence at The Islamia University of Bahawalpur
- 🧑🏫 Senior Executive Member of the AI Club — leading workshops and mentoring students
- 🏆 Competed in 10+ hackathons, including MIT Global AI Hackathons
- 📚 Trained through NAVTTC (1,600+ hrs) and the Pak AI Vision Group (CAI) program
- 🚀 Founded ZorceX to share practical AI knowledge and connect builders worldwide
- 📍 Based in Bahawalpur, Pakistan — working with clients globally
| Area | Focus |
|---|---|
| 🤖 AI Agents & LLMs | Tool-using agents, multi-agent orchestration, memory & reasoning patterns |
| 📚 RAG & Knowledge Systems | Hybrid retrieval, re-ranking, grounded answers, eval pipelines (RAGAS) |
| ⚙️ AI Automation | n8n / Make.com / Zapier workflows, MCP servers, multi-step agents + APIs |
| 🎙️ Voice AI | Real-time voice agents, streaming ASR/TTS, latency & turn-taking tuning |
| 🌐 Full-Stack AI Products | Next.js + FastAPI apps, auth, payments, investor-ready MVPs |
| ☁️ MLOps & Production | vLLM serving, observability (Langfuse), cost & latency optimization |
A full-spectrum ML practitioner — from mathematical foundations and classical models to transformer internals, LLM engineering, and aligned, production-grade systems.
|
|
|
|
RAGAS · DeepEval · Braintrust · Weights & Biases · MLflow · Langfuse · Phoenix / Arize · Guardrails AI · Llama Guard
flowchart LR
A["📥 Data Sources<br/>docs · APIs · streams"] --> B["🧹 Ingestion &<br/>Preprocessing"]
B --> C["🔢 Embeddings &<br/>Feature Store"]
C --> D[("🗄️ Vector DB<br/>pgvector · Pinecone")]
Q["🧑💻 User Query"] --> R["🔁 Hybrid Retrieval<br/>+ Re-ranking"]
D --> R
R --> O["🧠 LLM Orchestrator<br/>agents · tools · memory"]
O --> G["🛡️ Guardrails &<br/>Eval Hooks"]
G --> RESP["📤 Grounded Response"]
O -.-> M["📈 Observability<br/>Langfuse · traces · cost"]
G -.-> M
M -.-> FB["🔄 Feedback &<br/>Continuous Eval"]
FB -.-> B
style O fill:#6366f1,color:#fff
style D fill:#06b6d4,color:#fff
style G fill:#ef4444,color:#fff
style M fill:#8b5cf6,color:#fff
AI / LLM
Machine Learning & Deep Learning
Languages & Backend
Frontend
Data & Vector Stores
Automation & Voice
Infra & MLOps
| Project | Description | Stack |
|---|---|---|
| AI Credit Risk Advisor | Retrieval-augmented system for intelligent credit-risk insights | Python · LLMs · RAG · Vector DB |
| Image Model Fine-Tuning Pipeline | Automated SDXL / Flux.1 text-to-image fine-tuning toolchain | Python · Kohya_ss · Stable Diffusion |
| Autonomous Research Agents | Agents that plan, search, synthesize & return cited summaries | LangGraph · LangChain · RAG |
| MCP Servers & Tool Ecosystems | Internal APIs & actions exposed as composable agent tools | TypeScript · MCP · FastAPI |
| RAG for Large Codebases/Docs | Smart chunking, hybrid search & context packing at scale | Python · RAG · Vector DBs |
➡️ See more on my portfolio.
I write about AI system design, RAG, agents, and shipping production LLM applications on Medium and Substack.
I'm open to freelance work, startup builds, research collaboration, and scholarship recommendations.
