Mid/Senior AI LLM Engineer

Rekindly Studio; Warszawa; Remote; PLN 306,000 - 437,000; Full time

Benefits offered by this job

Remote work

20 days paid leave

Health insurance allowance

Equipment budget

Learning and development support

Annual team gathering

Job summary

Rekindly Studio, based in Europe, seeks engineers to fine-tune self-hosted LLMs powering AI conversations. You will train, evaluate, and deploy models, enhancing memory, multilingual support, and user experience. Collaboration spans AI engineering, product, web, and DevOps teams.

The role emphasizes production-ready workflows, experimentation, and responsible moderation. Remote, long-term B2B engagement with generous learning and tooling access.

Qualifications

• Proven hands-on experience fine-tuning open-weight LLMs and deploying them to production.
• Strong Python skills.
• Practical experience with PyTorch and Hugging Face.
• Experience with SFT, LoRA and QLoRA.
• Familiarity with preference-based training such as DPO or GRPO.
• Experience building training datasets and evaluation frameworks.

Responsibilities

• Train and fine-tune open-weight language models to improve conversational abilities.
• Build and refine training datasets for more accurate, natural responses.
• Create evaluation methods to measure model quality and identify improvements.
• Improve conversational memory, personality consistency and long-running dialogue.
• Develop multilingual capabilities and multi-character interactions.
• Investigate model failures and optimise training data, prompts and filters.
• Run GPU training experiments and troubleshoot issues.
• Deploy models to production, monitor performance and refine them.
• Collaborate with AI engineering, product, web development and DevOps teams.

Skills

Fine-tuning LLMs

Python

Production deployment

Model evaluation

Prompting and data analysis

GPU training debugging

Multilingual capabilities

Tools

PyTorch

Hugging Face

LoRA/QLoRA

Job description

Europe | Fully remote
B2B contract, long-term collaboration

The role

We are hiring for a company building AI-powered conversational products. You will fine-tune and improve self-hosted large language models that power personalised AI conversations, taking improvements from experimentation all the way to production.

What you will do

• Train and fine-tune open-weight language models to improve their conversational abilities
• Build and refine training datasets for more accurate, natural and consistent responses
• Create evaluation methods to measure model quality and spot areas to improve
• Improve conversational memory, personality consistency and long-running dialogue
• Develop multilingual capabilities and multi-character interactions
• Investigate model failures and optimise training data, prompts and response filters
• Run GPU training experiments and troubleshoot technical issues
• Deploy models to production, monitor performance and keep refining them
• Work closely with AI engineering, product, web development and DevOps teams

What you bring

• Proven hands-on experience fine-tuning open-weight LLMs and deploying them to production
• Strong Python skills
• Practical experience with PyTorch and Hugging Face (or similar tools)
• Experience with SFT, LoRA and QLoRA
• Familiarity with preference-based training such as DPO or GRPO
• Experience building training datasets and evaluation frameworks
• Hands-on GPU training and debugging
• Ability to analyse model behaviour and show measurable improvements
• Experience with consumer-facing AI products
• Comfortable working with uncensored models, explicit adult (NSFW) content and sensitive moderation material

What you get

• Fully remote work from anywhere in Europe
• Long-term collaboration on a B2B contract
• 20 days of paid leave
• $100 monthly health insurance allowance
• Equipment budget
• Access to AI development tools
• Learning and development support
• Annual team gathering
• Hands-on work with self-hosted LLMs used by real consumers