Open to opportunities

Andrei Zimin

@andreizimin

Data Scientist specializing in multimodal AI, reinforcement learning, and medical NLP—turning complex models into deployable products.

Zimbabwe

Message

What I'm looking for

I’m looking to build and ship production-grade ML/NLP systems end-to-end—owning model development, mentoring teams, and using RL, LLMs, and multimodal pipelines to deliver measurable product value, especially in healthcare and text processing.

I’m a Senior/Principal Data Scientist focused on bringing state-of-the-art AI—especially reinforcement learning and language models—into real products. I’ve led small data-science teams (3–4) and built systems from concept through MVP and pilot deployments. My work balances research depth with practical engineering, so models become reliable assistants for complex domains like healthcare.

In previous roles, I developed reinforcement learning-based automated drone control using PPO and trained YOLO models for object detection in aerial imagery, then built NER and NEL components for an AI-powered intelligence platform. More recently, I led technical development of a medical NLP product, designing pipelines for extracting, structuring, and standardizing clinical data from unstructured texts using LLMs and adapter fine-tuning. I also explore multimodal small language models, agentic systems, diffusion language models for throughput gains, and low-resource ASR/TTS workflows (including Whisper training and an Arabic ASR→LLM→TTS learning pipeline).

Experience

Work history, roles, and key accomplishments

Current

Principal Data Scientist

Current

Akcent

Feb 2024 - Present (2 years 5 months)

Led technical development of a medical NLP product from early-stage architecture to a working system, building pipelines to extract, structure, and standardize clinical data from unstructured texts using LLMs and adapter fine-tuning.

LLM Unstructured To Structured Pipelines Natural Language Processing

Senior Data Scientist

Aloha Mobile

Feb 2025 - Mar 2026 (1 year 1 month)

Trained multimodal small language models, enhanced an agent-based system into a browser environment, and developed classical time-series models for recommendation systems.

Agentic Systems Browser Agents Time Series Modeling Recommendation Systems

Data Scientist

Deep Knowledge Analytics

Jan 2023 - Dec 2023 (11 months)

Developed NER and NEL components for an AI intelligence platform that analyzed open-source data, calculated business metrics, highlighted company risks, and identified competitive strengths.

Named Entity Recognition Entity Linking NLP Data Analytics Market Research Open Source

Researcher

P.N. Lebedev Physical Institute

Oct 2021 - Jan 2023 (1 year 3 months)

Modeled collective behavior in multi-agent and complex systems, focusing on emergent dynamics and game-theoretic approaches relevant to reinforcement learning.

Multi Agent Systems Game Theory Reinforcement Learning

Data Scientist

Andrei Zimin

Oct 2020 - Oct 2021 (1 year)

Data Scientist at Andrei Zimin.

Data Scientist

NDA

Jan 2023 - Present (3 years 6 months)

Developed a reinforcement learning-based automated drone control system using PPO and trained YOLO models for object detection in aerial imagery.

Reinforcement Learning PPO YOLO Object Detection

Education

Degrees, certifications, and relevant coursework

Moscow Institute of Physics and Technology

Master of Science, Theoretical Physics

2018 - 2021

Researched a ternary choice model on graphs of arbitrary topology in the Potts model using Quantal Response Equilibrium. Completed an M.Sc. in Theoretical Physics.