09/24/2026 updated
60 % partially available
Senior Data Scientist & Applied AI Engineer | RAG, LLM, forecasting | Freelance, remote EU
Vilnius, Lithuania
Only remote
MSc Econometrics, University of Amsterdam, with honoursAbout me
Fifteen years building production systems — ten shipping machine learning: real-time ETA prediction from GPS streams, CTR prediction in a bidding stack, econometric demand forecasting, and a production RAG assistant. MSc Econometrics (Univ. of Amsterdam). Former Head of Engineering at Trafi.
MLOpsLangChainApplication Programming Interfaces (APIs)Data AnalysisEconometricsForecastingPython (Programming Language)Machine LearningNatural Language ProcessingPredictive ModellingFeature EngineeringData ScienceLarge Language ModelsDeep LearningGenerative AI
Data Science & Econometrics
Expertise in predictive modelling in production, demand and time-series forecasting, causal inference, A/B testing and experimentation, feature engineering, and model evaluation using Python, R, and SQL.
Applied AI & LLM Systems
Deep knowledge of agentic architectures (smolagents, LangGraph, LlamaIndex), RAG pipelines (LangChain, embeddings, vector retrieval), function calling, agent observability and evaluation, LLM APIs, PyTorch, fastai, and transformers.
Backend Engineering
Strong background in C#/.NET, Python, distributed and real-time systems, API and data-model design, ETL pipelines, AWS, payment/PSP integrations, and authentication.
Real-Time Transit Prediction
Building and productionising real-time transit prediction services from GPS streams, generating ETAs for consumer apps and public arrival displays.
Machine Learning for Ad-Tech
Development of machine-learning models for real-time click-through and conversion prediction using logistic regression and gradient boosting, including production ETL pipelines at bidding latency.
Econometric Demand Modelling
Construction of econometric demand models and forecasting tools adopted by 80+ analysts worldwide.
Payment & PSP Integration
End-to-end integration of virtual payment cards and PSP systems, including data compliance in regulated financial contexts and analytics pipelines for transaction insight.
Engineering Leadership & Architecture
Ownership of architecture direction for large-scale MaaS backends, mentoring engineers into senior and lead positions, and leading organisations of approximately 50 people through break-even and acquisition.
Deep Learning & Generative AI
Practical knowledge of CNNs, recurrent and transformer architectures, generative models, and recommender systems through formal training with fast.ai.
Expertise in predictive modelling in production, demand and time-series forecasting, causal inference, A/B testing and experimentation, feature engineering, and model evaluation using Python, R, and SQL.
Applied AI & LLM Systems
Deep knowledge of agentic architectures (smolagents, LangGraph, LlamaIndex), RAG pipelines (LangChain, embeddings, vector retrieval), function calling, agent observability and evaluation, LLM APIs, PyTorch, fastai, and transformers.
Backend Engineering
Strong background in C#/.NET, Python, distributed and real-time systems, API and data-model design, ETL pipelines, AWS, payment/PSP integrations, and authentication.
Real-Time Transit Prediction
Building and productionising real-time transit prediction services from GPS streams, generating ETAs for consumer apps and public arrival displays.
Machine Learning for Ad-Tech
Development of machine-learning models for real-time click-through and conversion prediction using logistic regression and gradient boosting, including production ETL pipelines at bidding latency.
Econometric Demand Modelling
Construction of econometric demand models and forecasting tools adopted by 80+ analysts worldwide.
Payment & PSP Integration
End-to-end integration of virtual payment cards and PSP systems, including data compliance in regulated financial contexts and analytics pipelines for transaction insight.
Engineering Leadership & Architecture
Ownership of architecture direction for large-scale MaaS backends, mentoring engineers into senior and lead positions, and leading organisations of approximately 50 people through break-even and acquisition.
Deep Learning & Generative AI
Practical knowledge of CNNs, recurrent and transformer architectures, generative models, and recommender systems through formal training with fast.ai.
Experience
Professional experience10+ years
Freelancer since2026
Languages
EnglishFluentLithuanianNative speaker
Project history
Owned architecture direction for the MaaS backend serving multimodal transport and payments across multiple countries. Built a RAG assistant with LangChain and LLM APIs over internal company documents. Led end-to-end integration of virtual payment cards in Switzerland. Introduced cost-allocation and engineering metrics supporting break-even profitability and a successful acquisition. Defined roles and mentored engineers into senior and lead positions across an organisation of approximately 50.
Built and productionised a real-time transit prediction service from GPS streams generating ETAs for consumer apps and public arrival displays in Jakarta. Designed and implemented core backend modules for multimodal transport, user management, authentication, and payments forming the foundation of the Jelbi platform in Berlin. Integrated dozens of mobility and payment service providers across cities. Grew the team from two engineers to eight and restructured it into three specialised teams.
Built Trafi's first machine-learning models predicting public-transit vehicle movements from real-time GPS data and shipped them into the mobile apps. Used observed GPS data to systematically improve the quality of static schedules, timetables, and routes.