I started by building products: deployed full-stack apps, real-time systems, mobile apps. Then, during an internship, a mentor showed me the AI side. I never looked back. Now I am an AI Engineering trainee (top 120 from 8,000+ applicants), building RL agents, completing an NLP capstone on Turkish dialogue, and I can still ship the backend.
I started as a full-stack developer: building real products, learning how systems work end-to-end. That foundation gave me something most ML engineers lack: I can ship.
During my internship at Çizgi Technology, I worked alongside an AI Research Engineer who redirected my entire trajectory. The problems I found most compelling were all on the AI side: autonomous systems, language models, intelligent pipelines.
I am a double major at Bahçeşehir University (Computer Engineering + Management Engineering), advancing through two competitive national AI programs, and looking for a team where I can contribute to real AI/ML work.
English C1 · Turkish Native · German A1
4-agent LangGraph pipeline with hybrid BM25 and ChromaDB retrieval using Reciprocal Rank Fusion. Claim-level hallucination detection in the ValidatorAgent and automatic retry loop on low-confidence answers. Evaluation layer tracking consistency, hallucination rate, cost and latency per trace, exposed through a FastAPI backend and Plotly Dash dashboard.
End-to-end ML pipeline on the NASA CMAPSS dataset to predict Remaining Useful Life of turbofan engines. Engineered temporal features over 5–20 cycle windows, benchmarked five models, achieved RMSE of 1.50 with Random Forest. Applied SHAP TreeExplainer for global and local interpretability and built a three-tier early warning system.
DQN and SAC algorithms across four simulation environments (Highway, Merge, Roundabout, Racetrack). Reward functions optimized for a zero-collision policy, high-speed navigation and strategic merging.
Real-time hospital communication platform, full-stack owner across all modules. React/Redux + FastAPI, Socket.IO, WebRTC, Gemini 2.5 AI assistant, live speech-to-text via AudioWorklet (48→16 kHz + VAD), JWT auth, AWS S3, encrypted PDF exports.
Full technical ownership in a cross-functional team. JWT auth, dynamic cart, admin dashboard. Deployed on Render & Vercel, iterated on UX alongside non-technical teammates.
Android chat application integrating OpenAI API (ChatGPT) for real-time conversation. Developed all activity classes: MainActivity, User, UI and logic layers, focusing on API bindings and data flow.
Accepted as an AI Software Developer for Baykar Technologies' 2026 Summer Internship Program before graduation.
Admitted to the AI Specialization Program, top 120 from 8,000+ applicants after intensive foundational screening (Data Structures, OS, DBMS). Focusing on Advanced Python, Cloud Systems, GPU Accelerators, and Parallel Programming. Capstone project under industry mentorship.
Selected as one of 1,500 participants from 31,700 applicants. National AI program supported by Ministry of Industry and Technology, Google Turkey, and T3 Foundation. Collaborating on AI-driven initiatives within a competitive ecosystem focusing on large-scale tech integration.
Built a full-fledged hospital communication panel under senior R&D mentorship. Real-time features with React, FastAPI, WebRTC, Socket.IO. Integrated Gemini 2.5 with live audio transcription (AudioWorklet, VAD). Delivered a production-ready demo.
Optimized Python/Django backend logic for faster data handling. Generated visual analytics dashboards with Pandas, NumPy and Matplotlib to monitor production efficiency. Used Watchdog library to track file changes and update database.
Double major: technical depth in CE (GPA 3.14) with systems thinking from ME (GPA 3.43). Capstone: Turkish dialogue NLP with BERTurk + Whisper. Core team member, BAU Management Engineering Club.
Open to AI/ML Engineering roles, Data Science positions, and interesting conversations. Based in Istanbul, open to hybrid and remote.
Send a Message