Senior Machine Learning - LLM Systems Engineer (f/m/x)
⚲ Białystok, Bydgoszcz, Gdańsk, Katowice, Kraków, Łódź, Lublin, Piła, Poznań, Rzeszów, Szczecin, Toruń, Warszawa, Wrocław
Do uzgodnienia
Wymagania
- Python
- C++
- CUDA
- Distributed ML Systems
- Triton (nice to have)
- Kubernetes (nice to have)
Opis stanowiska
Wymagania:
- Minimum 5 years of experience with LLM models and Machine Learning workloads
- Strong experience with machine learning in production environments
- Hands-on work with large-scale distributed ML systems
- Low-level optimization skills in C++, CUDA, or similar technologies
- Proven ability to analyze and improve ML system performance
- Solid Python skills with an understanding of system-level performance
- Very good English skills, both spoken and written
O firmie:
- Sii is the leading technology consulting, digital transformation, engineering, and business services vendor in Poland. We already employ more than 7 300 professionals and implement projects in a variety of industries for clients from many countries around the world. The Great Place to Work title, won 10 times in a row, proves that at Sii, we create a friendly work environment.
- Sii ensures that all employment-related decisions are made solely on the basis of qualifications and competencies. We are committed to equal treatment of all individuals, regardless of any protected characteristics, and promoting a diverse and inclusive work environment.
Zakres obowiązków:
- Scaling ML and LLM models across multi-node, multi-accelerator clusters
- Optimizing latency, throughput, and resource efficiency of ML workloads
- Designing and operating distributed training and inference systems
- Benchmarking and profiling AI workloads on real hardware platforms
- Identifying and resolving system-level performance bottlenecks
Oferujemy:
- Great Place to Work since 2015 - it’s thanks to feedback from our workers that we get this special title and constantly implement new ideas
- Employment stability - revenue of PLN 2.1BN, no debts, since 2006 on the market
- We share the profit with Workers - over PLN 76M has already been allocated for this aim since 2022
- Attractive benefits package - private healthcare, benefits cafeteria platform, car discounts and more
- Comfortable workplace – class A offices or remote work
- Dozens of fascinating projects for prestigious brands from all over the world – you can change them thanks to Job Changer application
- PLN 1 000 000 per year for your ideas - with this amount, we support the passions and voluntary actions of our workers
- Investment in your growth – meetups, webinars, training platform and technology blog – you choose
- Fantastic atmosphere created by all Sii Power People
- Minimum 5 years of experience with LLM models and Machine Learning workloads
- Strong experience with machine learning in production environments
- Hands-on work with large-scale distributed ML systems
- Low-level optimization skills in C++, CUDA, or similar technologies
- Proven ability to analyze and improve ML system performance
- Solid Python skills with an understanding of system-level performance
- Very good English skills, both spoken and written
O firmie:
- Sii is the leading technology consulting, digital transformation, engineering, and business services vendor in Poland. We already employ more than 7 300 professionals and implement projects in a variety of industries for clients from many countries around the world. The Great Place to Work title, won 10 times in a row, proves that at Sii, we create a friendly work environment.
- Sii ensures that all employment-related decisions are made solely on the basis of qualifications and competencies. We are committed to equal treatment of all individuals, regardless of any protected characteristics, and promoting a diverse and inclusive work environment.
Zakres obowiązków:
- Scaling ML and LLM models across multi-node, multi-accelerator clusters
- Optimizing latency, throughput, and resource efficiency of ML workloads
- Designing and operating distributed training and inference systems
- Benchmarking and profiling AI workloads on real hardware platforms
- Identifying and resolving system-level performance bottlenecks
Oferujemy:
- Great Place to Work since 2015 - it’s thanks to feedback from our workers that we get this special title and constantly implement new ideas
- Employment stability - revenue of PLN 2.1BN, no debts, since 2006 on the market
- We share the profit with Workers - over PLN 76M has already been allocated for this aim since 2022
- Attractive benefits package - private healthcare, benefits cafeteria platform, car discounts and more
- Comfortable workplace – class A offices or remote work
- Dozens of fascinating projects for prestigious brands from all over the world – you can change them thanks to Job Changer application
- PLN 1 000 000 per year for your ideas - with this amount, we support the passions and voluntary actions of our workers
- Investment in your growth – meetups, webinars, training platform and technology blog – you choose
- Fantastic atmosphere created by all Sii Power People
🔍 Dekoder Ogłoszenia
🔴
Minimum 5 years of experience with LLM models and Machine Learning workloads
Może oznaczać zarówno głębokie, specjalistyczne doświadczenie z LLM, jak i ogólne doświadczenie w ML, gdzie LLM są tylko jednym z wielu narzędzi.
🔴
Strong experience with machine learning in production environments
Może oznaczać zarówno budowanie i utrzymywanie złożonych systemów ML na produkcji, jak i jedynie wdrażanie gotowych modeli.
🔴
Hands-on work with large-scale distributed ML systems
Może oznaczać projektowanie i budowanie takich systemów od podstaw, jak i pracę z istniejącymi, gdzie zakres wpływu jest ograniczony.
🔴
Low-level optimization skills in C++, CUDA, or similar technologies
Wymaga to bardzo specjalistycznej wiedzy, która może być rzadko spotykana i niekoniecznie kluczowa dla wszystkich zadań związanych z LLM.
🔴
Proven ability to analyze and improve ML system performance
Może oznaczać głębokie analizy algorytmiczne i optymalizacje, ale też prostsze zadania związane z tuningiem parametrów.