Training an LLM stylist with expertise in fashion and style, capable of holding a dialogue, providing advice, and assisting with product navigation on Lamoda.
Conducting the full lifecycle of LLM fine-tuning: Continual pretraining, SFT, and Preference Optimization stages.
Developing RAG pipelines: full-text, semantic, and hybrid search of documents in the knowledge base.
Developing retrieval algorithms for various tasks: visual search, retrieving relevant documents, and finding suitable products/outfits.
Establishing the connection between the LLM stylist and outfit generation models (sets of products).
Developing internal benchmarks for evaluating language models, including LLM-as-judge approaches.
We expect:
5+ years of experience in ML, with 3+ years in Deep Learning.
Proficiency in Python and the ability to write complex SQL queries. Experience with Hadoop, Spark, Airflow, and Docker.
Experience in fine-tuning LLMs and building agent systems / RAG pipelines.
Experience in deploying ML models to production to solve business problems.
Knowledge of mathematical statistics, probability theory, algorithms, and data structures.
We offer:
Private health insurance covering dentistry, telemedicine, and mental health support programs.
Life insurance.
Monthly Lamoda promo codes: two 25% off, two 15% off, and a 40% off for your birthday.
Privileges and discounts program via BestBenefits.
Up to 15 fully paid sick days per year, covering 100% of your current salary.
Free legal consultations.
Additional days off for important life events.
Vibrant corporate events and abundant informal communication.
Development for everyone:
An experienced mentor assigned to every newcomer.
Training and opportunities to attend external conferences.
We help develop your personal brand, encourage speaking at conferences, and writing articles.
Performance Review twice a year to ensure systematic growth.
We organize meetups and demo days for knowledge sharing.
We immerse you in the world of fashion and introduce you to seasonal trends.