I am a Computer Engineering student at Gazi University—Faculty Valedictorian ranked 1st out of ~750 graduating students across the entire Faculty of Engineering with a 3.92/4.00 GPA (maintaining my rank as 1st among 120 students within the department). Alongside my studies, I served as a Reliability Engineering Trainee at Turkish Aerospace (TAI), where I bridged the gap between complex aerospace systems, safety-critical data, and autonomous software engineering.

I specialize in building highly optimized deep learning architectures, developing robust MLOps orchestration frameworks, and solving complex system-wide architectural bottlenecks. I don't just train models; I build the end-to-end cloud-native or edge infrastructure that keeps them running reliably in production.

• Languages & Core: Python, C#, SQL, Systems Engineering (FMEA/FMECA, FTA, ARP4754)
• MLOps & Cloud-Native: Docker, Container Orchestration, Redis, Terraform, AWS Architectures, MLflow
• AI & Deep Learning: PyTorch, TensorFlow, CNNs, Vision Transformers (ViT), Federated Learning
• NLP, GenAI & Backend: LangChain/AutoGen, GPT APIs, LLM Fine-Tuning & RAG, FastAPI, Flask

An enterprise-grade, cloud-native event-driven platform designed for real-time multi-modal streaming pipelines and autonomous agentic RAG.

• Simulates decoupled Kafka/Kinesis ingestion pipelines locally with an active checkpoint state-management loop to ensure fault-tolerance.
• Architected a self-healing MLOps loop that monitors embedding models for data/concept drift and triggers automated fine-tuning pipelines.
• Tech Stack: LangChain, AutoGen, Docker, Redis, Milvus, Terraform, AWS Blueprint (EKS, MSK, SageMaker).

A lightweight, Docker-first MLOps CLI tool for tabular data pipelines to detect feature and concept drift in production.

• Implements automated statistical testing (KS, PSI, ADWIN) against production streams to capture model degradation without manual intervention.
• Features a modular alert routing engine natively integrating Slack hooks and SMTP servers for automated system-health observability.
• Tech Stack: Python, Docker, NumPy, SciPy, River, Loguru.

A TÜBİTAK-supported high-dexterity prosthetic hand utilizing deep learning for real-time biological signal classification.

• Developed a Domain-Adversarial Neural Network (DANN) framework for sEMG gesture classification, hitting 87.7% validation accuracy via per-subject fine-tuning.
• Deployed the pipeline onto a Raspberry Pi 5, flashing inference latency down to an ultra-low 6 ms using TensorFlow Lite and XLA JIT compilation.
• Tech Stack: PyTorch, TensorFlow Lite, Python, Raspberry Pi 5, I2C / Hardware Integration.

A privacy-preserving, decentralized Federated Learning platform secured on top of the Ethereum blockchain.

• Utilizes custom Solidity smart contracts to orchestrate and enforce immutable auditability for local model weight distributions and updates.
• Mitigates single-point-of-failure vulnerabilities in classic FL architectures while keeping strict data governance intact.
• Tech Stack: Python, PyTorch, Solidity, Web3.py, Ethereum Ecosystem.

• Email: nizamfurkanegecan@gmail.com
• LinkedIn: linkedin.com/in/furkan-egecan-nizam
• GitHub Repositories: Explore more of my core infrastructure work right here on my profile!