AI Engineer with 8+ years of software engineering experience, currently building production agentic AI systems. I specialize in turning AI concepts into real, shipped, production systems — not prototypes — using:
- LLMs (OpenAI GPT, Anthropic Claude)
- RAG pipelines with evaluation-driven accuracy (RAGAS)
- Agent orchestration (LangGraph, LangChain)
- MCP (Model Context Protocol) tool integrations
- Cloud-native backend systems (AWS)
| Project | Result |
|---|---|
| Agentic RAG Support Assistant (Star Gallery) | ~40% faster ticket resolution — autonomous tool-use, zero human in the loop |
| Autonomous SRE Agent (Star Gallery) | ~70% less operational toil — fully autonomous DevOps operations |
| RAG Knowledge Assistant (Watermark Insights) | Response accuracy improved 71% → 89% via RAGAS-driven evaluation over 3 months |
| Turnitin AI Integration (Watermark Insights) | 60% faster evaluation across 50,000+ students |
- 🔍 Retrieval-Augmented Generation (RAG) systems — ingestion, embeddings, vector search, evaluation
- 🤖 Agentic AI workflows with autonomous tool-use (LangGraph, MCP)
- 🛡️ AI safety practices — guardrails, output validation, hallucination-risk mitigation
- 💬 AI chatbots with memory + summarization
- 🛒 AI-powered ecommerce intelligence systems
- ⚙️ High-performance backend APIs
- ☁️ Scalable AWS cloud architectures
- 🧬 Model fine-tuning (Hugging Face Transformers + LoRA — in progress: fraud detection)
🤖 AI Chatbot Backend Service — Open Source
Production-grade agentic RAG chatbot backend, built and maintained independently
- LangGraph-orchestrated pipeline: memory retrieval → semantic search → LLM generation → memory storage
- Dual-layer RAG with cosine similarity gating + MMR diversity ranking
- Multilingual support (English/Arabic)
- Redis-backed conversational memory, rate limiting
- Multi-LLM provider support (OpenAI, Anthropic, Groq)
- Fully Dockerized with CI pipeline and unit tests
Autonomous customer support agent with tool-use
- Multi-step agent workflow (LangChain, LangGraph, OpenAI, ChromaDB)
- Autonomously detects genuine customer issues and raises Jira tickets — no human in the loop
- ~40% faster ticket resolution across product policy, returns, and order queries
Self-operating DevOps automation agent
- LangGraph + Claude orchestration for SSL/TLS renewals, domain renewals, AWS health checks
- Fully autonomous — no human intervention required
- ~70% reduction in operational toil
Production RAG chatbot for institutional student-support documentation
- LangChain, LangGraph, OpenAI, vector search
- Reduced manual support queries by 35%
- RAGAS-driven evaluation improved response accuracy 71% → 89% over 3 months
- Agentic AI architectures & multi-agent orchestration
- RAG pipeline optimization & retrieval evaluation
- AI safety, guardrails & responsible AI practices
- Backend scalability & performance tuning
- Model fine-tuning & adaptation (LoRA)
- AWS cloud architecture
- Production AI systems that ship — not stay in notebooks
- 📧 Email: hasanfarhanascii@gmail.com
- 💼 LinkedIn: linkedin.com/in/hasandeveloper
- 🌐 Portfolio: https://www.hasanalicn.com
"Build systems that scale before they are needed — not after they break." Ship production, not prototypes. Measure everything.