machine learning engineer

Enfint; Berlin; Vor Ort; EUR 90.000 - 150.000; Vollzeit

Benefits dieser Stelle

Stock options

Hybrid work model

Wellness program

Career growth

Zusammenfassung

Bolt is seeking an experienced machine learning professional to design, train, and deploy ML/LLM models for catalogue enrichment, moderation, and categorisation at scale. You will own production accuracy, latency, and cost while transitioning to fine-tuned open-weight models and building robust data pipelines.

You will build agentic AI systems for catalogue automation, including multi-step workflows, tool use, guardrails, and evaluation harnesses, and oversee offline benchmarks, A/B tests, and

Qualifikationen

• Proven experience building and shipping production ML systems at scale, ideally for consumer-facing products at a technology company.
• Experience taking LLM-based systems to production, including prompt and model iteration, evaluation, guardrails, observability, and cost management
• Hands-on experience fine-tuning and serving open-weight models, including LoRA/QLoRA, SFT, preference optimisation, training data development, and production serving
• Deep expertise in NLP or recommendation systems, with models that improved a business metric
• Mastery of Python and SQL, clean code practices, and production experience with a deep learning framework such as PyTorch, TensorFlow, JAX, or Triton
• Experience with modern ML tooling and cloud infrastructure, including AWS, SageMaker, Airflow, and Docker
• Strong product sense and proactive ownership, with the ability to turn ambiguous problems into measurable ML solutions and drive them from discovery to production impact
• Nice to have: experience in consumer-facing product environments

Aufgaben

• Design, train, and deploy ML/LLM models for catalogue enrichment, moderation, and categorisation at scale.
• Drive the transition from frontier API models to fine-tuned open-weight models by building data curation and fine-tuning pipelines.
• Design and build agentic AI systems for catalogue automation, including multi-step workflows, tool use, guardrails, and evaluation harnesses.
• Build evaluation and experimentation infrastructure, including offline benchmarks and A/B tests linked to business metrics.
• Own serving and cost management for self-hosted models, including quantisation, throughput tuning, GPU utilisation, and build-versus-buy decisions.
• Collaborate with Software Engineers, Data Scientists, Product Managers, and the ML Platform team to productionise and monitor ML solutions.

Jobbeschreibung

Описание:

Bolt is a technology company operating mobility and delivery services across 50+ countries and 850+ cities, serving over 200 million customers through a platform supported by 4.5+ million partners and 4,000+ employees. Its mission is to make cities for people, not cars.

Задачи:

Design, train, and deploy ML/LLM models for catalogue enrichment, moderation, and categorisation at scale, owning production accuracy, latency, and cost; Drive the transition from frontier API models to fine-tuned open-weight models by building data curation and fine-tuning pipelines, comparing quality, and taking selected models into production; Design and build agentic AI systems for catalogue automation, including multi-step workflows, tool use, guardrails, and evaluation harnesses; Build evaluation and experimentation infrastructure, including offline benchmarks, regression suites, LLM-as-judge pipelines, and A/B tests linked to business metrics; Own serving and cost management for self-hosted models, including quantisation, throughput tuning, GPU utilisation, and build-versus-buy decisions; Collaborate with Software Engineers, Data Scientists, Product Managers, and the ML Platform team to productionise and monitor ML solutions.

Требования:

• Proven experience building and shipping production ML systems at scale, ideally for consumer-facing products at a technology company
• Experience taking LLM-based systems to production, including prompt and model iteration, evaluation, guardrails, observability, and cost management
• Hands‑on experience fine‑tuning and serving open-weight models, including LoRA/QLoRA, SFT, preference optimisation, training data development, and production serving
• Deep expertise in NLP or recommendation systems, with models that improved a business metric
• Mastery of Python and SQL, clean code practices, and production experience with a deep learning framework such as PyTorch, TensorFlow, JAX, or Triton
• Experience with modern ML tooling and cloud infrastructure, including AWS, SageMaker, Airflow, and Docker
• Strong product sense and proactive ownership, with the ability to turn ambiguous problems into measurable ML solutions and drive them from discovery to production impact
• Nice to have: experience in consumer-facing product environments

Условия:

• Rewarding salary and stock option package
• Hybrid working with a minimum of 3 days in the office each week
• Wellness perks supporting physical and mental health
• Career opportunities and professional growth
• Work on products serving millions of customers and partners across 850+ cities in 50+ countries
• Some perks may differ depending on location and role