data scientist · medellín, colombia
Models that show their work.
I'm Sebastian — Master's in Data Science & Analytics (EAFIT). I build market-trend models, large-scale NLP pipelines on Spark, and LLM agents that run in production — and I publish the metrics.
fig. 1 — momentum gradient descent on a non-convex loss surface
scroll to descend ↓
§ 1 — selected projects
Case studies, with the numbers attached
§ 2 — toolbox
Skills, with receipts
modeling
- XGBoost · CatBoost
- Scikit-learn
- TensorFlow · PyTorch
- Time-series validation
→ used in forex trend prediction
nlp & big data
- Spark NLP · PySpark
- AWS EMR · Hadoop
- Topic modeling (LDA)
- MLflow tracking
→ used in bank complaints NLP
llm engineering
- RAG pipelines
- OpenAI · Groq / Llama
- Flask · MongoDB
- Telegram · WooCommerce APIs
→ used in sales chatbot
analysis & ops
- Pandas · NumPy · SQL
- Hypothesis testing
- Power BI
- Docker · AWS
→ everywhere above
§ 3 — education
Background
-
2023 — 2025
M.Sc. Data Science & Analytics — EAFIT University
Machine learning, deep learning, and NLP. GPA 4.6 / 5.0. Thesis on ML models for forex trend prediction.
-
2022
Data Scientist Certificate — Microsoft · IDATA · EAFIT
Certification in data science and machine learning.
-
2014 — 2020
B.Eng. Civil Engineering — Universidad Cooperativa de Colombia
Thesis: optimization of substation foundations.
§ 4 — contact
Let's work together
Open to data science and ML engineering roles, and to interesting problems in between. Email is fastest — or find me on LinkedIn and GitHub.
