Greater Istanbul
Senior AI & Platform Engineer with 7+ years of experience building production-grade AI systems, LLM-powered applications, and scalable distributed platforms in enterprise environments. I specialize in designing and delivering AI infrastructure, agentic AI systems, and MLOps platforms that enable organizations to reliably move from AI prototypes to production-grade deployments. My focus is on building systems that are scalable, observable, governable, and model-agnostic. My core expertise lies in AI platform engineering, agent orchestration, and LLM systems design, including building autonomous agents, multi-step reasoning pipelines, and tool-augmented LLM architectures. I have extensive hands-on experience with frameworks such as LangChain and LlamaIndex, and I design advanced RAG systems using vector databases like Qdrant, Pinecone, and pgvector. On the infrastructure side, I build backend systems, microservice architectures, and asynchronous event-driven pipelines, deployed at scale on Kubernetes in cloud-native environments. I focus heavily on system reliability, performance optimization, and distributed systems design for high-traffic workloads. I also design and implement MLOps and AI lifecycle systems, including CI/CD pipelines for ML/LLM applications, model and prompt evaluation frameworks, and production monitoring using Prometheus, Grafana, and ELK stack. This includes observability, tracing, logging, cost monitoring, and drift detection for AI systems. Core Expertise AI Platform Engineering & Infrastructure Agentic AI Systems & LLM Orchestration MLOps, CI/CD for ML/LLM Applications Distributed Systems & Cloud-Native Architectures RAG Pipelines (Qdrant, Pinecone, pgvector) Kubernetes, Microservices, Async Systems Observability, Monitoring & Reliability Engineering Evaluation & Regression Testing for AI Systems I am passionate about bridging the gap between cutting-edge AI research and production-ready enterprise systems, enabling organizations to deploy AI safely, reliably, and at scale.
As part of the core engineering team, I work on the infrastructure layer that makes agentic AI reliable, governable, and shippable in real production environments. I design and implement LLM-agnostic agent runtimes, control planes, and evaluation systems that enable AI agents to operate reliably, deterministically, and at scale in enterprise environments. Technical Responsibilities Architecting agent execution runtimes (sandboxed execution, async workflows, streaming, retries, timeouts) Building agent orchestration and control planes (Plan–Execute–Reflect loops, task graphs, state & memory management) Designing evaluation & regression testing frameworks for LLM-based agents and prompts Implementing observability pipelines (distributed tracing, metrics, logs, cost & latency monitoring) Developing policy enforcement and governance layers (tool restrictions, budget controls, auditability) Creating model-agnostic adapters for multiple LLM, embedding, and tool providers Engineering CI/CD pipelines and deployment workflows for agent systems Solving failure handling, determinism, and reliability challenges in agent-based architectures Core Focus Areas Agent Reliability Engineering · Agent Evaluation & Testing · Control Plane Architecture · Distributed Systems · Cloud-Native Platforms · MLOps for Agentic AI · System Safety · Enterprise Governance Tech Stack & Keywords Agentic AI · LLM Systems · AI Infrastructure · Platform Engineering · Distributed Systems · Python · Go · Kubernetes · Docker · Async Processing · Event-Driven Architectures · Observability (Tracing, Metrics, Logging) · CI/CD · Cloud Platforms · Model-Agnostic AI
• Led the design and implementation of scalable distributed systems and backend services across Yaska Group subsidiaries. • Owned the integration of AI/ML and LLM into production environments, enhancing operational efficiency. • Developed intelligent automation platforms and agentic AI systems, driving innovation within the organization. • Spearheaded Docker & Kubernetes-based cloud-native deployments, optimizing CI/CD pipelines for multiple teams.
AI MLOps Engineer | Specializing in Agentic AI, LLMs & VLM Infrastructure At a premium AI company, I own the MLOps lifecycle for products built on AGI, LLMs, VLMs, and Agentic AI. My core focus is engineering the systems that allow autonomous agents to operate reliably and at scale. My Responsibilities & Technical Competencies: Agentic AI Development & Deployment: Designing, scaling, and monitoring the full lifecycle of autonomous agents capable of multi-step reasoning, tool use, and goal completion. Integrating with frameworks like LangChain or LlamaIndex for agent orchestration. RAG Pipeline Development: Designing and maintaining end-to-end Retrieval-Augmented Generation systems, from vector database integration (e.g., Pinecone, Chroma) to semantic search, providing agents with real-time, contextual knowledge. Model Optimization: Applying advanced quantization techniques (e.g., QLoRA, GGUF) to reduce model size and latency, enabling efficient agent deployment on a wider range of hardware. Infrastructure & Automation: Managing infrastructure as code (IaC) using Docker, Kubernetes, and Terraform. Automating workflows with tools like MLflow, Kubeflow, Airflow and for seamless business process integration. Observability: Establishing centralized monitoring and alerting infrastructure for agent decision-making, tool usage, system metrics, and logs (ELK Stack, Prometheus/Grafana). My goal is to build the engineering foundation that ensures even the most advanced AI agents operate seamlessly, reliably, and effectively in production. I'm always eager to learn new technologies and solve complex problems at the intersection of MLOps and Agentic AI. Don't hesitate to connect to expand my professional network and explore collaboration opportunities. #MLOps #LLMOps #Kubernetes #Docker #CI/CD #MachineLearning #AI #LLM #VLM #GenerativeAI #Python #AWS #GCP #DevOps #DeepLearning #RAG #Quantization #n8n #AgenticAI #AIAgents #LangChain
- **LLM Fine-Tuning & RAG Implementation:** Fine-tuned **GPT-4 and LLaMA** models using **Hugging Face Transformers** for a customer support chatbot, achieving a **15% improvement in response accuracy**. Implemented **Retrieval-Augmented Generation (RAG)** with **Qdrant** vector database, reducing latency by **20%**. - **Model Optimization:** Applied **LoRA and 4-bit quantization** to reduce model size by **40%** while maintaining **95% of the original performance**. - **Scalable AI API Deployment:** Developed **RESTful APIs** using **Flask/FastAPI**, containerized with **Docker**, and deployed on **Kubernetes** clusters on GCP. Handled **10,000+ requests per second** with **99.9% uptime**. - **CI/CD Pipeline Optimization:** Reduced deployment time by **40%** by implementing **GitHub Actions** and **Jenkins** for automated testing and deployment.
Developed computer vision-based navigation systems for autonomous mobile robots using OpenCV, PyTorch, and ROS (Noetic/Humble), achieving 95% obstacle detection accuracy. These systems served as the foundation for critical perception modules like dirt detection and floor classification. • Leveraged open-source Hugging Face multimodal models to build NLP-based command recognition systems, enabling robots to interpret and execute complex verbal commands with potential integration into Vision-Language-Action (VLA) models. Optimized inference speed by 30% through the application of CUDA acceleration and model distillation techniques for edge computing devices. Led R&D initiatives for sensor fusion and deep learning algorithm integration into robotic hardware platforms; integrated Lidar, RealSense, and RGB camera data using advanced data processing algorithms for robust environmental perception, floor analysis, and dirt density mapping. Designed robust backend communication layers to handle telemetry data between robots and central control systems using Python-based microservices. Computer Vision (OpenCV, Image Classification, Object Detection, Semantic Segmentation) Machine Learning (PyTorch, TensorFlow, Scikit-learn) Robotics (ROS, ROS2, Lidar, RealSense SDK, Sensor Fusion) Model Optimization (ONNX, TensorRT, Quantization, CUDA Acceleration)