10/29/2026 updated
100 % available
Senior AI/ML Engineer | Python, LLMs & Data Science
Utrecht, Netherlands
Only remote
PhD in Science (Biophysics), Complutense University of Madrid, 2007–2011About me
I build Python AI applications, document-based chatbots and automated workflows using OpenAI, AWS Bedrock, FastAPI and SQL. With a PhD and experience in AI, banking and research, I deliver tested solutions, clear documentation and practical handover.
AI AgentsRAGOpenAIBedrockLangChainAmazon Web ServicesData AnalysisPython (Programming Language)Machine LearningNatural Language ProcessingPredictive ModellingSQL DatabasesFeature EngineeringPrompt EngineeringFastAPI
Applied AI and LLM Engineering
LLM applications including agent orchestration, tool calling, structured outputs, prompt engineering, RAG, embeddings, semantic search, multilingual conversations, synthetic research personas and multimodal analysis. Development of custom agent workflows and shared Python workflow frameworks for production AI products.
Machine Learning and Predictive Modelling
Supervised classification, regression, clustering, imbalanced learning, feature engineering, cross-validation, LightGBM, Optuna, scikit-learn, Bayesian analysis, stochastic simulation and time-series forecasting. Applications include campaign response modelling and loan-booking prediction.
Python Data Engineering and Cloud Infrastructure
Proficient use of pandas, NumPy, SQL, MLflow, Prefect, FastAPI, Streamlit, Pydantic and pytest for building reusable data pipelines and analytical workflows. Cloud and integration experience with AWS Lambda, Amazon Bedrock, S3, DynamoDB, Kinesis, CloudWatch, IAM, Secrets Manager, Docker, Git and CI/CD workflows.
LLM Evaluation and Prompt Optimisation
Evaluation workflows for generated text, classifiers and complete application behaviour. Use of OpenAI Evals, MLflow tracing, custom requirement-following judges, labelled test sets, model and provider comparisons, critical-error analysis and prompt-optimisation experiments.
Natural Language Processing and Text Analysis
Development of NLP features including survey-response coding, text quality assessment, knowledge retrieval, multilingual chatbot support and named entity extraction. Experience with retrieval-augmented generation, semantic search and structured output validation.
Quantitative Research and Scientific Computing
Quantitative research including stochastic simulation, time-series analysis, image analysis and Bayesian modelling applied to biophysics and cell biology. Use of MATLAB, PyTorch and scientific image analysis tools alongside Python-based workflows.
Investment Research and Algorithmic Trading
Design of a connected Python trading platform covering stock selection, forecasting, portfolio management, options research and commodity futures. Integration of LLM-based research agents with deterministic execution controls, broker APIs (Alpaca, SnapTrade, Deribit, IBKR) and risk management workflows.
Document Digitisation and OCR
Building of Python workflows that convert scanned PDFs and handwritten documents into editable Word and PDF outputs. Use of vision-language models for text transcription, structured layout reconstruction, handwriting review and named entity extraction with annotation preparation for Label Studio.
Reporting, Forecasting and Stakeholder Communication
Automated contact-history and conversion reporting with pandas and Excel. Application of ARIMA forecasting for planning, structured experiment design with CRM and digital colleagues, and communication of KPIs and findings to commercial and insurance teams.
Experience
Professional experience5 - 10 years
Freelancer since2026
Languages
EnglishFluentSpanishNative speaker
Project history
Develops NLP and generative AI features with product, research and engineering colleagues. Contributions include adaptive survey follow-ups, agentic interview research, multilingual customer support chatbot, image/audio/video understanding, synthetic personas, survey coding knowledge retrieval, custom agent workflows, structured data mapping, privacy checks, LLM evaluation and prompt improvement, and operational deployment support.
Built a personal assistant supporting text and voice conversation using the OpenAI Realtime API, with tool calling, file inspection, asynchronous audio queues and configurable model selection.
Built a personal application that listens to the microphone and displays translated text during a conversation. Uses local speech recognition through Whisper MLX or Parakeet and Google Translate via googletrans.