#!/usr/bin/env python3
class TalhaSari:
def __init__(self):
self.name = "Talha Sarı"
self.career_arc = [
"medical AI → computer vision → backend engineering",
"full-stack RAG products → enterprise multi-agent systems",
"MCP-based developer tooling",
]
self.current_focus = [
"reliable coding-agent workflows",
"context engineering",
"harness engineering",
"evaluations",
"deterministic controls",
"improvement loops",
]- Early RAG products: As SafeVideo AI's first and Founding AI Engineer, I worked across system design and hands-on implementation for early RAG and user-facing AI products.
- Enterprise multi-agent product: As Jetlink's first AI engineer, I built a multi-agent product across AI, backend, and frontend, brought it into production, and it is still used by large enterprises.
- Internal MCP tooling: At Garanti BBVA Teknoloji, I build internal MCP-based tools that integrate AI into software-engineering workflows.
- Generative AI benchmark: Co-authored a JMIR Medical Informatics study comparing six LLMs on medical-paper understanding through a RAG benchmark with expert reference answers.
- National first place: Won first place with the MedicAI team in TEKNOFEST 2022's university-and-above computer-vision disease-detection category.
- Languages: Python, TypeScript, Go
- AI systems: PyTorch, Hugging Face, computer vision, LLMs, RAG, AI agents, MCP, evaluations
- Backend & data: FastAPI, Node.js, Supabase, MongoDB, vector databases
- Infrastructure: Docker, AWS, Git, CI/CD, Linux