Hi there, I'm Talha 👋

---

Full-Stack AI Engineer | Architecting & Scaling LLM Systems

#!/usr/bin/python
# -*- coding: utf-8 -*-

class TalhaSari:
    def __init__(self):
        self.name = "Talha SARI"
        self.role = "AI Engineer"
        self.current_focus = [
            "Autonomous Multi-Agent Architectures",
            "High-Performance RAG Systems",
            "Full-Stack AI Platform Development"
        ]

    def get_key_achievements(self):
        return {
            "Open Source Impact": {
                "Project": "AutoLLM",
                "Achievement": "Engineered a framework simplifying RAG, achieving 80,000+ downloads."
            },
            "Product Innovation": {
                "Achievement": "Transformed a simple chatbot into a Multi Agent AI system with dynamic tool use.",
                "Impact": "Eliminated costly manual integrations, reducing deployment time from weeks to hours."
            }
        }

    def __str__(self):
        return f"Thanks for visiting! I'm focused on building impactful AI. Let's connect."

me = TalhaSari()
print(me.get_key_achievements()["Open Source Impact"])

🚀 Highlights

• Platform Architect: Spearheaded the ground-up migration of all AI functionalities from a C# monolith to a high-performance, containerized Python microservices platform for a system serving 80+ B2B customers.
• Open Source Creator: Engineered and launched AutoLLM, a popular open-source framework that simplifies RAG development, achieving over 80,000 downloads and 1,000+ GitHub stars.
• Performance Optimizer: Re-architected a RAG retrieval engine using Milvus and an HNSW index, slashing search latency by ~99% (from 500ms to <5ms).
• Product Innovator: Transformed a legacy chatbot into a multi-agent system, replacing a costly manual webhook process with a no-code API framework and cutting client deployment times from weeks to hours.
• AI Researcher: Led R&D on Vision-Language Models (VLMs) for automated video analysis and built an LLM-as-a-Judge framework to programmatically evaluate model faithfulness and quality.
• National Award Winner: Led my team to 1st Place out of 127 teams in the National AI in Healthcare Competition at TEKNOFEST 2022.

---

🛠️ My Tech Stack

• Languages: Python, JavaScript, SQL
• AI/ML: PyTorch, Computer Vision, Hugging Face, NLP, Prompt Engineering, LLMs, RAG, MCP, AI Agents
• Databases: MongoDB, Milvus, LanceDB, Supabase
• Cloud & DevOps: Azure, AWS, Docker, FastAPI, Git, CI/CD, Linux

---

📫 Let's Connect

I'm always open to discussing new projects, innovative ideas, or opportunities in AI. Feel free to reach out!

• LinkedIn: linkedin.com/in/talhasari/
• Email: talhasarit@gmail.com