Hire the Best Data Scientists
in Turkey

Applied Scientist with 14+ years of experience designing statistical and machine learning models that directly impact revenue, risk, and operational decisions. Strong background in probabilistic modeling, forecasting, ranking, and optimization, with hands-on experience deploying models into production across aviation, banking, telecom, and e-commerce. PhD in Statistics with published research on entropy-based learning and attention-driven feature selection.

Are you looking to integrate custom LLM Agents, ChatGPT/Claude/Gemini, or highly accurate RAG pipelines into your systems without the operational complexity? I build deterministic, scalable, and cost-optimized AI solutions that drive real business value—from initial concept to full AWS deployment. I am a Full-Stack AI/ML Engineer who bridges the gap between cutting-edge Artificial Intelligence and robust software engineering. I don't just train models in notebooks; I conceptualize, design, develop, test, refine, and deploy AI systems that integrate seamlessly into your existing software infrastructure. My approach focuses on solving the biggest bottlenecks in modern AI: reducing API costs, eliminating latency, preventing hallucinations, and lowering operational complexity. Whether you need a custom LangGraph multi-agent system, an intelligent CRM integration, or a scalable backend API, I build solutions with scalability, speed, and maintainability in mind. CORE EXPERTISE & HIGH-DEMAND SOLUTIONS I BUILD: LLM Integrations & Advanced Agentic Workflows • API Integrations: Deep expertise in OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini, and open-source models (Qwen, Llama, Mistral). • Agent Architectures: Building complex, multi-agent systems using LangGraph and custom frameworks. • Cost & Speed Optimization: Implementing model routing, response caching, and prompt engineering to keep API costs down while maximizing speed. • Deterministic AI: Forcing LLMs to produce structured, predictable outputs using Pydantic, JSON schemas, and function calling. • Smart Feature Extraction: Offloading tasks from expensive LLMs to efficient algorithms whenever possible to save you money and ensure accuracy. RAG Pipelines & Intelligent Data Processing • Enterprise RAG: Building Retrieval-Augmented Generation systems (document ingestion, hybrid retrieval, response generation) with extreme accuracy and consistency. • Vector Search: Implementing advanced search algorithms using Vector Databases (Pinecone, Chroma, etc.). • Document Parsing & OCR: Extracting actionable data from PDFs, Word docs, Excel, CSVs, and integrating them into your AI systems. AI Automations & Third-Party Integrations • Workflow Automation: Building no-code/low-code pipelines using n8n, Make, and Zapier to create end-to-end AI-powered workflows. • Platform Integrations: Connecting AI agents to Slack, ClickUp, WhatsApp, Twilio, Shopify, Amazon, Monday, Notion, Airtable, HubSpot, Salesforce, Zapier, Make, Google Workspace, Microsoft Teams, Discord, Telegram and existing CRMs to automate business management. • Webhooks & Custom APIs: Developing secure webhooks to route data flawlessly between platforms. Voice AI & Computer Vision • Speech-to-Text & Text-to-Speech: Integrating state-of-the-art voice pipelines (WhisperX, Faster-Whisper, Deepgram, ElevenLabs, Retell, Pyannote for speaker diarization). • Computer Vision: Developing comprehensive Deep Learning solutions including Object Detection (YOLO architectures), Image Classification, Facial Recognition, and Semantic Segmentation (DeepLabV3+). Expertise in PyTorch/Torchvision frameworks and building robust image processing pipelines with OpenCV. Data Science, Predictive ML & LLMOps • Predictive Analytics: Beyond GenAI, I utilize the Python data stack (Pandas, NumPy, Scikit-learn, XGBoost) for deep data analysis, forecasting, and classification on structured business data. • Security & Monitoring: Ensuring enterprise data privacy by deploying air-gapped local LLMs (Ollama, vLLM) and using LLMOps tools (LangSmith) to track agent behavior, rigorously evaluate RAG accuracy, and prevent prompt injections. Full-Stack Deployment, MLOps & Cloud Infrastructure • Backend Development: Writing powerful Python scripts and building ultra-fast APIs with Python FastAPI. • Cloud & DB: Deploying on AWS (EC2, S3, RDS, Lambda) with PostgreSQL and MongoDB architectures. • Containerization & Optimization: Dockerizing applications, utilizing ONNX and Quantization techniques to run heavy ML models on CPU for maximum cost efficiency. • Frontend/Dashboards: Creating interactive UIs with Streamlit for quick demonstrations and internal tools. PROVEN TRACK RECORD: As a graduate of Software Engineering and a highly prestigious, government-backed AI & Technology Academy, I thrive in high-pressure, results-oriented environments. My portfolio includes fully deployed systems such as: • AI-Powered Background Removers with Google OAuth and AWS deployment. • Fine-tuned NLP & RAG systems generating automated test queries. • Full-scale App Review Analyzers using BERTopic and AWS RDS. • Local, offline AI Agents with real-time web search capabilities. Let's Build Something Impactful. I am highly communicative, proactive, and dedicated to solving your specific business bottlenecks. Send me a message, and let's discuss how we can turn your data and ideas into a deployed, revenue-generating AI product.

AI automation engineer. I build AI agents, RAG pipelines and business process automation that a company actually runs on, in Python with Claude, Claude Code, MCP servers, LangChain and LangGraph. $260,000 earned on Upwork across 125 completed projects, 5,229 hours billed, 100% Job Success Score, Top Rated Plus. Most AI work dies in the gap between a model that works in a demo and a process that works on a Monday morning. Closing that gap is the whole job. Clients I have built for include Trukkin (logistics), LGFG Fashion House (fashion), Ultrassure (insurance), Genvision (climate tech), Bowery Boost (performance marketing), Moor Marketing (lead generation) and Ta3swim (ecommerce). Istanbul, UTC+3. I overlap 09:00 to 13:00 US Eastern every weekday and answer inside 4 hours outside that window. You work with me directly. I write the code myself. WHAT I BUILD AI Agents and LLM Systems in Production Single and multi-agent systems, LangGraph and LangChain orchestration, Claude API and OpenAI API integration, MCP servers and connectors, Claude Code and Claude Cowork setups, function calling, human approval gates wherever a mistake is expensive, plus evaluation and guardrails so you can tell whether the agent is actually working. AI Automation and System Integration Business process automation and CRM automation across the tools your team already uses. Two way syncs between systems that were never designed to talk to each other. APIs, webhooks, and the payload changes that quietly break things overnight. ZOHO CRM, internal tools, email, databases and data flows across your stack. Python orchestration first; no-code tools only as human-in-the-loop glue where that is the right call. Fixing and Hardening What You Already Have A workflow that ran fine for months and then stopped. An agent that answers confidently and is wrong. A prototype that works on one laptop with the API keys sitting in a file. I diagnose it, fix it, and leave it on a footing where it keeps running: logging, error handling, retries, version control, permissions. This is most of my current work. Three of my active contracts are improving and extending systems that were already in production, including a weekly ML agents program running since February 2025. RAG Over Your Own Documents Retrieval augmented generation that stays accurate as the document set grows, with source citations so any answer can be checked. Chunking, embeddings, Pinecone, ChromaDB, GraphRAG and evaluation. When a first attempt disappoints, the problem is almost always retrieval, not the model. Also AI chatbots for support and sales grounded in your own data, with escalation to a human when confidence is low. SaaS and MVP Development End to end AI powered SaaS builds. Python and FastAPI backends, authentication, billing, dashboards, cloud deployment on AWS, GCP or Azure. From a blank repository to a live product with paying users. Data Pipelines and Scraping Large scale web scraping, ETL, cleaning and enrichment. The unglamorous layer that decides whether an AI system is actually useful. RESULTS BY INDUSTRY Real Estate: 30+ automation projects (Zillow, CoStar), AI powered house price calculators, ZOHO CRM integrations. Digital Marketing: AI SEO researchers, content generators, YouTube analytics agents, full blog production pipelines. Healthcare and SaaS: AI chatbots, personalized email systems, lead generation platforms taken from zero to live products. Sales and Go to Market: AI sales agents, lead research and qualification, outreach personalization at scale. CURRENTLY BUILDING A weekly ML agents program on a single production system, running since February 2025. A RAG orchestration and prompt engineering layer on top of an existing LLM product. An AI content pipeline in daily production use. Long engagements rather than one off scripts, which is where the 100% Job Success Score comes from. TECH STACK LLMs and Agents: Claude, Claude Code, Claude API, OpenAI GPT, Gemini, MCP servers and connectors, agentic and multi-agent architectures, LangChain, LangGraph, function calling, prompt engineering, evaluation and guardrails RAG and Data: RAG pipelines, GraphRAG, embeddings, vector databases, Pinecone, ChromaDB, PostgreSQL, ETL, large scale scraping Backend: Python, FastAPI, Node.js, REST APIs, webhooks, Docker Automation and Integration: API integration, business process automation, CRM automation, ZOHO CRM, workflow automation Cloud: AWS, GCP, Azure HOW I WORK I ask about the business before I write code. I tell you when something should not be automated. I document what I build in plain English, so your team can own it after I am gone. Send me the process you want covered and I will tell you what it can own, what stays with your team, and how long it takes.

I build LLM agent systems that hold up in production — LangGraph orchestration, RAG over messy real-world data, and the evaluation layers that keep them reliable. Recent work: a multi-agent platform managing 13 interdependent workstreams with automatic dependency invalidation and a reviewer agent gating output; a retrieval layer combining keyword and semantic search with cross-encoder reranking and MCP tool calling against live APIs; a LangGraph conversational agent with deterministic guardrails living outside the model and an adversarial evaluation harness; a healthcare claims pipeline combining deterministic rules with fine-tuned models under HIPAA, where auditability decided the architecture. Also a full-duplex voice assistant over an ERP backend — streaming ASR, semantic endpointing, barge-in, WebRTC. Most of my time goes to routing, evaluation and failure handling rather than model calls. Integrating an LLM API takes a week; making it reliable enough that someone acts on the output is the job. Stack: Python, LangGraph, FastAPI, PostgreSQL + pgvector, Docker, Anthropic/OpenAI/Gemini, self-hosted inference via vLLM and Ollama. 7+ years, MSc in Artificial Intelligence, two published papers (ACL, SemEval).

As a highly experienced data scientist with over five years of hands-on expertise, I specialize in developing and deploying advanced machine learning solutions. My core competencies lie in Natural Language Processing (NLP), Generative AI, and Computer Vision. I have a proven track record of building robust, production-level models that solve complex business challenges and drive innovation.

If you need healthcare data pipelines, analytics systems, AI-powered workflows, or biostatistical analysis, I can help you develop a reliable solution tailored to your project. I combine healthcare data engineering and biostatistics with clinical research experience demonstrated by 62 medical articles indexed in PubMed. I CAN HELP YOU WITH: Data Engineering & AI • Build and maintain production ETL/ELT pipelines • Develop and manage Apache Airflow and Google Cloud Composer workflows • Design BigQuery pipelines, SQL transformations, and analytics systems • Integrate EHR platforms, third-party APIs, and external data sources • Automate manual file transfers and system-to-system data movement • Extract structured clinical data from PDFs, scanned documents, and laboratory reports • Build controlled AI-agent workflows for healthcare data • Automate KPI dashboards and recurring reports • Develop machine-learning solutions for healthcare analytics Biostatistics & Clinical Research • Analyze clinical, observational, and survey data using Python or IBM SPSS • Prepare descriptive statistics and hypothesis tests • Perform linear and logistic regression analyses • Conduct survival analysis • Apply parametric methods, including t-tests, ANOVA, ANCOVA, MANOVA, and repeated-measures ANOVA • Apply nonparametric methods, including Mann–Whitney U, Kruskal–Wallis, Wilcoxon signed-rank, and Friedman tests • Interpret statistical findings in their clinical and scientific context • Create publication-ready tables, figures, and statistical summaries • Review statistical methods and results sections for manuscripts • Communicate findings clearly to clinicians, researchers, and nontechnical stakeholders Send me a brief description of your project. For statistical work, you may also share the research question, dataset structure, or a de-identified sample. I can provide a complimentary preliminary scope review and suggest practical solution options. If one of those approaches fits your needs, we can then discuss the implementation details. WHY WORK WITH ME? My background combines production data engineering with genuine clinical and research experience. Before moving into technology, I worked as a medical doctor and researcher who published 62 medical articles indexed in PubMed. This experience enables me to understand both the technical architecture and the real-world meaning of healthcare data. I know that a statistically correct result must also be clinically meaningful and that a technically functional system must be reliable, maintainable, and appropriate for its healthcare environment. I can work directly with medical terminology, research questions, clinical outcomes, and study documentation without requiring clients to translate the clinical context into purely technical language. Selected experience: • Built AI-assisted workflows that extract laboratory data from unstructured PDF reports and load structured results into BigQuery • Developed and maintained healthcare data pipelines integrating multiple EHR and external data sources • Optimized Google Cloud Composer workloads to improve reliability and control infrastructure costs • Developed machine-learning and anomaly-detection pipelines for IoT monitoring data • Conducted biostatistical analyses for clinical research and peer-reviewed publications Core technologies: Python, SQL, Apache Airflow, Google Cloud Composer, BigQuery, dbt, GCP, Docker, Pandas, NumPy, scikit-learn, XGBoost, LangChain, Agno, Kafka, REST APIs and IBM SPSS.