Looking for a job

Santiago Fada

@santiagofada

AI Research Engineer at Eclypsium building reinforcement learning systems that cut costly decompilation timeouts by 45%.

Argentina

Message

At Eclypsium, I built and shipped a production reinforcement learning library that predicts costly decompilation timeouts and decides whether to run them. It cut timeouts by 45% with minimal coverage loss.

I also developed an unsupervised anomaly detection pipeline with PyOD to flag binaries that differ from groups of similar files, and prototyped LLM-based detection of vulnerabilities and other security issues.

At Mercado Libre, I deployed a cold-start recommendation model across seven countries, improving offline F1-score by 20%. I validated it with online A/B tests on CTR before rollout and integrated machine learning workflows into CI/CD pipelines.

My other work includes computer vision models for detection, classification, tracking, and segmentation, as well as time-series forecasting and RAG-based solutions for document retrieval and question-answering. I have an Applied Mathematics background and support practical sessions as a Teaching Assistant at Universidad Nacional de Córdoba (FaMAF).

Experience

Work history, roles, and key accomplishments

Current

AI Research Engineer

Current

Eclypsium

Feb 2026 - Present (9 months)

Built and shipped a production reinforcement learning library that predicts decompilation timeouts, cutting them by 45%. Developed an unsupervised anomaly detection pipeline and prototyped LLM-based vulnerability detection.

Reinforcement Learning Thompson Sampling Deep Q Learning PyOD Anomaly Detection Isolation Forest ECOD COPOD LLM Python

Current

Machine Learning and AI Solutions

Current

Freelance

Jan 2026 - Present (10 months)

Developed time series forecasting workflows for food service businesses and built RAG-based solutions for document retrieval and question-answering.

Time series Forecasting SARIMA Prophet RAG Lang Chain LlamaIndex Pinecone Embeddings Vector Search LLM

Current

Teaching Assistant

Current

Universidad Nacional de Córdoba (FaMAF)

Mar 2025 - Present (1 year 8 months)

Teaching assistant in the Applied Mathematics program, supporting practical sessions and answering student questions in courses such as Modeling and Simulation and Introduction to Machine Learning.

Teaching Applied Mathematics Modeling and Simulation Machine Learning

Machine Learning Developer

Mercado Libre

Sep 2024 - Feb 2026 (1 year 5 months)

Deployed a cold-start recommendation model across 7 countries, improving offline F1-score by 20%. Integrated ML workflows into CI/CD pipelines and supported blue-green and staged deployments.

Recommendation Systems Cold Start Strategies F1 Score A B Testing CI CD MLFlow Data Drift Model Versioning Python

Data Scientist

UNXDigital / Grupo Prominente

Apr 2024 - Sep 2024 (5 months)

Designed and trained computer vision models for detection, classification, tracking, and segmentation tasks in industry, finance, and logistics. Worked end-to-end from data preprocessing to model evaluation and deployment support.

Computer Vision Object Detection Classification Trackings Segmentation data preprocessing Model Evaluation Deployment

Data Analyst (Power BI)

Esolutions

Feb 2023 - Feb 2024 (1 year)

Performed data transformation, modeling, KPI definition, and business intelligence analysis. Built dashboards and reports to support strategic decision-making.

Power BI data transformation Data Modeling KPI Definition Business Intelligence Dashboarding Reporting