Senior ML Engineer (JetBrains Research)

JetBrains

JetBrains is a global software company that creates intelligent tools for software developers and teams. Since 2000, we have built products that help developers work more productively, write higher-quality code, and stay focused on solving real problems. The JetBrains Research team is looking for a Senior Research Engineer to work on Code World Models: models that learn how software systems behave, change, execute, and interact with developer tools. This role is focused on model pre-training and mid-training for code-centric foundation models. You will work on data, training pipelines, evaluation, and experiments that improve how models understand programs, repositories, execution, tests, and software engineering workflows. IN THIS ROLE, YOU WILL: * Design and run pre-training, continued pre-training, and mid-training experiments for code models. * Build and improve data pipelines for large-scale model training, including filtering, deduplication, mixture design, and dataset quality checks. * Work with code corpora, repositories, tests, execution traces, and synthetic data. * Develop evaluations for complex repository-level code reasoning tasks. * Collaborate with researchers and engineers working on ML for code and AI developer tools. WE’LL BE HAPPY TO HAVE YOU ON OUR TEAM IF YOU: * Have hands-on experience with model pre-training, continued training, or mid-training. * Have strong engineering skills in Python and experience with modern ML frameworks. * Understand large-scale ML training workflows, including data processing, distributed training, checkpointing, evaluation, experiment tracking, and debugging. * Have experience working with large datasets and care about data quality, contamination, sampling, and reproducibility. * Have a background in NLP, ML for software engineering, or a similar domain. * Enjoy working on research problems with high uncertainty and turning ideas into working experiments. IT WOULD BE A PLUS IF YOU: * Have experience training or adapting models for code generation, code understanding, software agents, program repair, test generation, or repository-level reasoning. * Have worked with execution-based data, such as unit tests, traces, logs, compiler feedback, runtime states, or sandboxed code execution. * Have experience with large-scale distributed training of models with 70B+ parameters. * Understand evaluation challenges for code models, including benchmark contamination, flaky tests, execution-based scoring, and long-horizon task evaluation. * Have contributed to ML infrastructure, open-source projects, or research systems. #LI-KP1 We are an equal opportunity employer We know great ideas can come from anyone, anywhere. That’s why we do our best to create an open and inclusive workplace – one that welcomes everyone regardless of their background, identity, religion, age, accessibility needs, or orientation. We process the data provided in your job application in accordance with the Recruitment Privacy Policy.

JetBrains

At JetBrains, code is our passion. Ever since we started, back in 2000, we've been striving to make the strongest, most effective developer tools on earth. Today, AI-powered coding agents are becoming a core part of how developers write Kotlin – and we want to make sure they write it well. The Kotlin AI Value Stream team is responsible for how AI agents understand, generate, and improve Kotlin code across all platforms: Android, Kotlin Multiplatform, server-side, web, desktop, and others. We build the evaluation infrastructure, error analysis tools, and post-training pipelines that measure and improve agent behavior on real Kotlin developer tasks. As a Research Engineer on this team, you'll own the end-to-end loop: Analyze how agents fail on Kotlin → build evals that capture those failures → research and implement methods to fix them → measure the improvement. Your work will directly shape how millions of developers experience Kotlin through AI coding agents. AS PART OF OUR TEAM, YOU WILL: Build tools for agentic error analysis * Design and implement tooling to systematically capture, classify, and analyse errors that AI coding agents make when generating Kotlin code. * Build observability pipelines over agentic traces – mining patterns from agent sessions in JetBrains IDEs, Junie, Claude Code, Cursor, and other coding agents. Build evaluation pipelines * Design, implement, and maintain evaluation pipelines that measure Kotlin code generation quality across dimensions, including correctness, idiomaticity, build success, framework usage, and test coverage. * Build simulation environments where coding agents can be measured on realistic Kotlin developer tasks – from greenfield KMP projects and Gradle dependency management to migrating Spring applications from Java to Kotlin. * Own evaluation infrastructure: metrics, experiment tracking, automated regression checks, and reproducible benchmarking. Research methods for improving agent and model behavior on Kotlin * Experiment with post-training techniques (SFT, DPO, GRPO) to improve how models handle Kotlin-specific patterns, idioms, and frameworks. * Investigate context engineering approaches: CLAUDE.md/AGENTS.md files, compiler-as-verifier feedback loops, Kotlin LSP integration, and MCP-based tooling. * Run experiments to measure impact: A/B comparisons, benchmark suites, and before/after analyses on real codebases. * Collaborate with model providers (Anthropic, OpenAI, and Google) to translate Kotlin-specific findings into model improvements. Build public Kotlin benchmarks * Design and build open-source benchmarks that measure AI coding agent performance on Kotlin tasks and eventually become the standard reference for the ecosystem. * Create task datasets covering the breadth of Kotlin usage: the server side (Spring, Ktor), multiplatform projects (KMP), build systems (Gradle), Android, library development, and others. * Include both mined real-world tasks and carefully designed synthetic tasks that test specific Kotlin capabilities. * Maintain and evolve benchmarks as models improve, ensuring they remain challenging, relevant, and contamination-resistant. WE'LL BE HAPPY TO HAVE YOU ON BOARD IF YOU HAVE: * Hands-on experience building evaluation or analysis pipelines for LLMs or AI coding agents in a research or production setting. * Strong Python engineering skills (at least three years), with the ability to write clean, maintainable code in data-heavy and ML-adjacent codebases. * Experience with data analysis at scale: querying large datasets (SQL/Athena), building data pipelines, and performing statistical analysis of experimental results. * The ability to own projects end to end – from identifying a problem in agent traces to designing an eval, running experiments, and shipping a fix. * A product-aware mindset: You care about how agents are actually used by developers and can translate real failure modes into evaluation and training work. * Familiarity with Kotlin or a strong willingness to develop deep Kotlin expertise (you'll be living in Kotlin codebases daily). OUR IDEAL CANDIDATE WOULD ALSO HAVE EXPERIENCE WITH: * Post-training LLMs: SFT, RLHF, DPO, GRPO – either hands-on training or designing the data and reward pipelines that feed into training. * Modern deep learning frameworks (PyTorch) and LLM training stacks (TRL, verl, Megatron, or similar). * AI agent development: tool-using agents, multi-step coding workflows, agentic frameworks. * Evaluation frameworks and tools: Inspect AI, Promptfoo, LM-evaluation-harness, or custom eval pipelines. * Experiment tracking and observability: Weights & Biases, MLflow, Langfuse, or similar. * The Kotlin ecosystem: Android, Gradle, KMP, Spring, Ktor – with an understanding of the developer workflows that agents need to support. * Contributing to or maintaining open-source projects, especially benchmarks or evaluation tools. Don't check every box? That's okay – if you're excited about this work and bring strong fundamentals, we'd love to hear from you. We're happy to talk and provide the training you need to grow into the role. WHY JOIN JETBRAINS?  * Strong base salary. We offer competitive pay that reflects your skills and experience. * Flexible work location. Enjoy the freedom to work from home or from the office. * Remote work. Spend up to 30 days per year working remotely from abroad. * Extra time off. More days to relax, recharge, and do the things you love. * Medical insurance allowance. Enjoy peace of mind for you and your family * Learning and development opportunities. Access to conferences, courses, and language classes. * Relocation support. We help make your move as smooth and stress-free as possible.  * Language classes. Pick up the local language or sharpen your English skills. * Fuel your day. Enjoy a hot meal or receive a lunch allowance on workdays. * Mental health support. To help you feel your best, we provide easy access to professional mental health services. * Sports benefit. Enjoy an on-site gym or sports club stipend. * Internal events. Join company-wide celebrations and team gatherings. *Some benefits may vary depending on location. #LI-KP1 We are an equal opportunity employer We know great ideas can come from anyone, anywhere. That’s why we do our best to create an open and inclusive workplace – one that welcomes everyone regardless of their background, identity, religion, age, accessibility needs, or orientation. We process the data provided in your job application in accordance with the Recruitment Privacy Policy.

JetBrains

At JetBrains, we build developer tools used by millions of engineers. The AI for Code team works on the next generation of coding agents and agentic workflows: systems that can understand codebases, plan and execute multi-step tasks, collaborate with developers, and ship reliable results inside real development environments. We are looking for a Staff/Senior AI Engineer to join the team and support these efforts. This role is for someone who can take our internal coding models, such as Mellum2, as well as open-weight models, and turn them into production-ready coding agents for our users. You’ll work on model training and fine-tuning, context engineering, tool use, evaluation, feedback loops, and product integration. This is not research in isolation – you’ll build systems that are used by tens of thousands of developers. WHAT YOU’LL DO: * Build production-ready coding agents and agentic workflows for real developer tasks inside JetBrains products. * Turn promising model capabilities into dependable product behavior through prompt design, context construction, fine-tuning, instruction-tuning, or other post-training techniques where appropriate. * Design and improve the agent loop itself, including tool use, execution strategy, safeguards, and task completion quality. * Create evaluation suites and quality infrastructure for agent behavior, including online and offline evaluations, regression checks, failure analysis, and release criteria. * Build feedback loops from real usage, using logs, user signals, and edge cases to improve data, evaluations, and agent behavior. * Work with both hosted frontier APIs and self-hosted or open-weight models, making pragmatic decisions about where each model belongs based on capability, latency, reliability, privacy, and cost. * Collaborate closely with product managers, software engineers, ML engineers, and researchers to ship features end to end. * Help define the technical direction for future work, especially in ambiguous areas where we need strong judgment rather than a prewritten playbook. WHAT WE’RE LOOKING FOR: * Strong software engineering fundamentals and a track record of shipping complex systems to production. * Hands-on experience building LLM-powered products, coding agents, or other AI systems. * Experience improving model behavior through systematic iteration, whether via prompting, context engineering, fine-tuning, preference optimization, or broader post-training methods. * Practical experience with evaluation and benchmarking for LLM systems, including defining task-grounded success metrics and catching regressions. * Experience working from noisy real-world signals rather than only from clean benchmark datasets. * Good judgment about trade-offs between model quality, latency, reliability, privacy, and cost. * Confidence working with ambiguity and taking ownership of a direction over multiple iterations. * Strong communication skills and the ability to align engineering and product decisions. WHAT SUCCESS LOOKS LIKE IN THE FIRST YEAR: * You ship one or more agent capabilities that users can rely on for meaningful work, not just demos. * You establish better evaluation coverage and clearer release criteria for agent behavior. * You help the team build a repeatable loop from idea to shipped capability: prototype, evaluate, learn from usage, improve, and scale. WHY JOIN US? You’ll help define what practical, trustworthy AI for software development looks like in real products. You’ll work on challenging problems at the boundary of model capability and product reality, with the freedom to stay hands-on and the scope to influence how the next generation of JetBrains AI systems is built.  And: * Strong base salary. We offer competitive pay that reflects your skills and experience. * Flexible work location. Enjoy the freedom to work from home or from the office. * Remote work. Spend up to 30 days per year working remotely from abroad. * Extra time off. More days to relax, recharge, and do the things you love. * Medical insurance allowance. Enjoy peace of mind for you and your family * Learning and development opportunities. Access to conferences, courses, and language classes. * Relocation support. We help make your move as smooth and stress-free as possible.  * Language classes. Pick up the local language or sharpen your English skills. * Fuel your day. Enjoy a hot meal or receive a lunch allowance on workdays. * Mental health support. To help you feel your best, we provide easy access to professional mental health services. * Sports benefit. Enjoy an on-site gym or sports club stipend. * Internal events. Join company-wide celebrations and team gatherings. *Some benefits may vary depending on location. #LI-KP1 We are an equal opportunity employer We know great ideas can come from anyone, anywhere. That’s why we do our best to create an open and inclusive workplace – one that welcomes everyone regardless of their background, identity, religion, age, accessibility needs, or orientation. We process the data provided in your job application in accordance with the Recruitment Privacy Policy.