Senior AI Engineer

About the Role

We're looking for a Senior AI Engineer with a PhD in engineering or a related field to design, build and ship production-grade generative AI systems. You'll work on LLM-powered products used by real customers: multi-agent workflows that automate complex business processes, retrieval-augmented assistants built on large internal knowledge bases, and evaluation pipelines that keep them reliable.

This role sits where applied research meets engineering. You'll bring scientific rigor to problems that don't have a known answer yet. You'll turn recent research into working systems, design experiments that show what actually works, and develop new methods when existing ones fall short. You'll also carry ideas all the way to production, making them scalable, observable and cost-efficient.

You'll work closely with product, data and platform teams, and you'll help shape our technical and research direction as the field changes fast. Current projects include [e.g., an agentic customer-operations assistant, a document intelligence platform, and an internal copilot for engineering teams].

Key Responsibilities

• Lead applied research: Find open problems in our GenAI systems, form hypotheses, and design and run controlled experiments with sound methodology and statistical analysis.
• Turn research into products: Track the latest work in literature, reproduce promising papers, and adapt them to our products and constraints.
• Design and build agentic systems: Build single-agent and multi-agent architectures (tool use, planning, memory, human-in-the-loop checkpoints) using frameworks like LangGraph, Agent SDKs, or custom orchestration. Invent new approaches to agent reasoning, planning and reliability where off-the-shelf patterns fall short.
• Build RAG pipelines: Own ingestion, chunking, embedding, hybrid search, reranking and grounding strategies across vector databases.
• Integrate and optimize LLMs: Work with commercial APIs and open-weight models. Choose models based on quality, latency and cost. Apply prompt engineering, structured outputs and fine-tuning where it adds value.
• Set up evaluation and quality: Create evaluation frameworks with offline test sets, LLM-as-judge, regression suites and online metrics. Define new metrics and benchmarks for our domain, and check their validity, reliability and statistical significance.
• Productionize and operate: Deploy AI services with solid APIs, observability (tracing, token and cost monitoring), caching, rate limiting and fallbacks using Docker and Kubernetes.
• Build in safety and governance: Add guardrails, PII handling, prompt-injection defenses and access controls. Make sure systems meet data privacy and compliance requirements.
• Lead technically: Mentor engineers, review code and designs, write technical docs, and help decide when to build, buy or wait. Coach the team on research methodology and experimental design.
• Work with stakeholders: Turn business problems into clear AI solutions and research questions with measurable success criteria and explain trade-offs to non-technical stakeholders.

Qualifications

Required

• A PhD in Computer Engineering, Electrical & Electronics Engineering, Computer Science, Industrial Engineering, or a closely related field, with a thesis in machine learning, NLP, information retrieval, optimization, or a related area
• 5+ years of combined doctoral research and industry experience, including 2+ years building LLM or GenAI applications that shipped to production
• Strong Python skills (Fast API, async programming, testing) and solid software engineering fundamentals
• Hands-on experience with RAG, embeddings and vector search in production
• Experience building agentic workflows: tool/function calling, multi-step reasoning, state management and error recovery
• A deep, practical understanding of LLM behavior, limitations and failure modes, and how to evaluate them systematically
• Experience with cloud platforms, containerization and CI/CD
• Clear written and spoken communication in English

Nice to Have

• Fine-tuning or distilling models and serving them with vLLM, TGI or similar
• Research experience in reinforcement learning, multi-agent systems, planning or reasoning
• Experience with Model Context Protocol (MCP) or building tool ecosystems for agents
• Knowledge of LLM observability tools
• Background in NLP, information retrieval or classical ML
• Open-source contributions, publications or public AI projects
• Experience with funded research projects (e.g., TÜBİTAK, Horizon Europe) or university–industry collaborations

Am I Right for the Team?

You'll probably enjoy working with us if:

• You ship. A working system in users' hands matters more to you than a perfect demo in a notebook.
• You connect research and product. You love reading papers, and you get even more satisfaction from seeing an idea make a measurable difference for real users.
• You measure before you believe. You've been burned by "it looked great in testing," so you build evals first. Controlled experiments and ablations are second nature to you.
• You're pragmatic about the stack. Sometimes the right answer is a well-written prompt, sometimes a fine-tuned model, sometimes a new method you design yourself, and sometimes no LLM at all.
• You stay curious. You keep up with new models and research, and you can tell hype from real progress.
• You like ownership. You're comfortable with ambiguity and would rather shape the problem than just get a ticket.
• You lift others up. You share knowledge freely, give thoughtful code reviews, and help the team grow.

This might not be the right fit if you want pure research with no production responsibility, or if you prefer tightly defined, fully specified tasks.