Data Scientist — LLM/Agentic Systems

About the Position

EOS Data Analytics is a product company that transforms satellite images into real-world solutions. Our platforms help farmers in over 100 countries optimize their yields, forest services track deforestation, and energy companies find resources-all thanks to data from Earth's orbit. Our main office is in the USA, with R&D centers in Kyiv and Dnipro.

EOS Data Analytics is looking for an AI/LLM Engineer to design and implement agentic systems, automation pipelines based on LLM, and AI-driven workflows integrated with satellite Earth observation and business processes.

Responsibilities

• Designing and implementing multi-agent architectures for automating data processing and analysis.
• Creating RAG pipelines and tools based on LLM for internal products and customer-oriented solutions.
• Developing evaluation environments to measure the quality of agent systems.
• Integrating LLM API ensembles (Anthropic, OpenAI, DeepSeek, etc.) into workflows.
• Researching and prototyping new agent patterns (worker swarms, orchestrators, tool use).
• Collaborating with various teams to explain and coordinate developed solutions and practical automation tasks.

Requirements

• 3+ years of experience in ML/AI engineering.
• 5+ years of experience working with Python (adhering to PEP standards and OOP principles).
• Deep understanding of LLM APIs and prompt engineering (including tool use, function calling, structured outputs).
• Experience with agent frameworks: LangFamily, GSDK, CrewAI, or custom orchestrators.
• Knowledge of FastAPI, asynchronous Python, and REST API design.
• Experience designing RAG pipelines: data ingestion, chunking, searching, retrieval, reranking, and optimizing prompts and costs.
• Understanding of evaluation patterns for multi-model systems (LLM-as-judge, scoreboarding, RAGAS, HITL).
• Experience with Docker and AWS (EC2, S3).
• Experience with system setup (audit logging and monitoring).

Nice to Have

• Experience integrating MCP (Model Context Protocol) / A2A.
• Knowledge of agent system security patterns: risk classification, deliberative gates, immutable audit logs, guardrails.
• Experience with SQLite or graph-based world models for managing agent states.
• Experience using AWS SageMaker or Lambda for serverless LLM inference.
• Knowledge of vector databases (Pinecone, Weaviate, pgvector, etc.).
• Experience in R&D, ability to confidently work in iterative prototyping environments.

We Offer

• Health: Health insurance, 20 paid vacation days, 100% paid sick leave up to 30 days per year, sports compensation and sessions with a psychologist or coach, flexible start time.
• Development: Course compensation, conference attendance, language learning, corporate library, online courses, expert lectures, internal meetups-as a participant or speaker.
• Comfort: Coworking compensation where there is no office, gifts for important life events, financial support in critical situations.
• Outside Work: Team-building activities, corporate events, social initiatives-helping Ukraine, supporting special children, environmental projects.

EOS DA is changing how humanity recognizes Earth from space. Join us!

Welcome! I'm Vlad. I'd be happy to chat with you about engines, architecture, and high loads.

@vladislav_hruba

Send your resume