Senior Machine Learning Engineer

Working mode: 📍Hybrid in Krakow (1-2 days/week office)

The organization is searching for a Senior Machine Learning Engineer to work on LLM-based systems, focusing on integration, end-to-end processes, and production readiness. This position combines performance improvement (prefill/decode, throughput, latency) with hands-on experience integrating components throughout the AI platform.

A substantial portion of the job is creating end-to-end integration flows and tests, notably for token generation pipelines and system orchestration, as well as contributing to intelligent system behavior like hardware selection and execution methods. The position also include creating system-level logic in Python for multi-tenant management, caching methods, and service lifecycle management throughout the platform.

*Not Data Science position*

Responsibilities:

• Build and maintain end-to-end integration flows across the AI inference pipeline (serving, orchestration, APIs, and infrastructure)
• Design, implement, and optimize LLM inference workflows, including prefill and decode stages
• Improve system performance with focus on throughput, latency, and interactivity
• Write production-grade components in Python and integrate them into the broader system
• Contribute to system-level logic such as smart hardware selection and execution strategies
• Integrate models (open source and custom), services, and APIs into cohesive, reliable end-to-end application pipelines

Requirements

• 4+ years of experience in software engineering or machine learning engineering
• Strong proficiency in Python
• Strong experience with LLM inference systems and performance optimization (must have)
• Hands-on experience with system integration and end-to-end workflows
• Experience with inference frameworks such as vLLM, TensorRT, SGLang etc
• Experience working with GPU/accelerator-based systems

Preferred Qualifications

• Hands-on experience with Dynamo and LLM-D for LLM inference and serving
• Familiarity with Kubernetes and cloud environments