Building ML systems that work in production and publish in ACL.
Currently pursuing an M.Sc. in Information Engineering at Universidad de los Andes (GPA 4.63/5.0), with research at the Conecta-TE lab under the Academic Vice-Rectory. My work focuses on educational analytics, NLP, and machine learning applied to real-world systems.
With industry experience at Habi and Duppla, I have built production-grade ML pipelines for real estate price prediction, content forecasting, and automated property valuation. My research has been peer-reviewed and published at ACL 2025 (Mumbai) and Springer/ICICT 2025 (London).
ML / DL
Data & Cloud
MLOps & Infra
BI & Viz
Universidad de los Andes — Conecta-TE
Top-ranked Colombian university · Educational analytics research lab
Universidad de los Andes
Top-ranked Colombian university · Graduate AI program
Duppla
Colombian PropTech startup · Automated real estate valuation
Habi
Colombian PropTech unicorn · Real estate marketplace
Retrieval-augmented generation system using chunking, embeddings, and vector storage paired with an LLM for accurate document question answering. Deployed on GCP with REST API.
Hybrid extractive-abstractive pipeline using PACSUM + LLM prompting, outperforming fine-tuned Legal-Pegasus and BART models. Ranked 7th internationally at JUST-NLP 2025, published in ACL proceedings.
CNN + BiLSTM model for early detection of 6 harmful brain patterns from EEG signals (Kaggle HMS dataset), achieving 63.98% accuracy (F1: 0.6282) using bipolar montage preprocessing.
Content-based recommendation engine using sentence embeddings and cosine similarity for semantic matching between user prompts and book summaries. Deployed on GCP.
E. G. Soto Parada, C. M. Muñoz Almeida, D. Cuevas Alba
Proc. JUST-NLP Workshop @ ACL 2025, pp. 155–161 · Mumbai, India
E. G. Soto Parada, G. A. Montoya, C. Lozano-Garzón
Proc. ICICT 2025 · London, UKISBN 978-981-96-6441-2
Universidad de los Andes
Focus: ML, NLP, Educational Analytics · Graduate Research Assistant at Conecta-TE
Jan 2025 – Present
Universidad de los Andes
Thesis in Machine Learning graded 5.0/5.0 · TA: Modeling & Simulation, Data Structures & Algorithms
Feb 2020 – Nov 2024
Open to research collaborations, ML engineering roles, and interesting problems.