Data Science / ML Engineer (Universal Role)

Vacancy: Data Science / ML Engineer (Universal Role) Company: IMS City: Saint Petersburg Format: Remote Employment: full-time / part-time Salary Range: Negotiable (1500-2000/hour, depends on candidate's level, fixed payment is negotiable) Employment Type: Civil Law Contract (GPX), Individual Entrepreneur (IP), Self-employed. NDA signing is mandatory, and we have high confidentiality requirements.

Responsibilities • Develop the product's AI-core: from processing user requests to final responses and result visualization. • Design and improve retrieval/pipeline logic for multi-sources (Qdrant, MemGraph, PostgreSQL, non-relational sources). • Develop the LLM agent core (Query Planner loop): request classification, subtask decomposition, tool calls, assessment of data sufficiency. • Work with embeddings, semantic search, and output quality (relevance, deduplication, filtering, ranking). • Participate in optimizing LLM inference (Qwen family, LiteLLM proxy), request costs, and latency. • Set up and improve data/ML pipelines for processing documents, multimodal content, and preparing data for search/graph. • Interact with the backend team (FastAPI/WebSocket), participate in designing contracts and service integrations. • Participate in developing technical solutions for scaling and performance (Redis caching, load, stability).

Requirements • 2–5 years of experience in Data Science / ML Engineering (Middle / Middle+/Senior). • Proficient Python and experience in production development of ML/DS services. • Practical experience with LLM/NLP: prompting, call orchestration, response quality assessment, pipeline optimization. • Experience with vector search and embeddings (preferably Qdrant or similar Vector DB). • Understanding of working with graph and relational DBs (MemGraph/Neo4j, PostgreSQL) in the context of AI search and analytics. • Experience in designing APIs/integrations (REST; understanding of WebSocket and service interaction). • Ability to formulate hypotheses, conduct experiments, and make reasoned technical decisions.

Will be a big plus • Experience building RAG/agentic systems in production. • Experience optimizing LLM models and inference infrastructure (latency/cost/quality trade-off). • Knowledge of Redis caching and session-aware caching approaches. • Experience with multimodal scenarios (text + documents/media). • Experience formalizing service contracts (including gRPC/YAML/OpenAPI). • Experience building data processing and classification pipelines.

We offer • Work on a technologically complex product at the intersection of LLM, search, and graph data. • Opportunity to influence the AI part of the architecture and key technical decisions. • Flexible remote work format. • Competitive salary based on experience level. • Strong team, minimal bureaucracy, and live communication.

Contact: recruiter contact