ML ENGINEER · BOGOTÁ, COLOMBIA
Shipping production LLM, RAG, and agentic systems on AWS.
Machine Learning Engineer with hands-on experience shipping production LLM, RAG, and agentic systems on AWS for document-intensive enterprise pipelines. MSc in Biomedical Engineering with research on multimodal data collection, computer vision, and deep learning.
Comfortable working end-to-end across ML backend, serverless infrastructure (AWS Lambda, Step Functions, SAM), and full-stack delivery. Focused on turning frontier AI methods — LLM evaluation, retrieval, and human-in-the-loop workflows — into reliable, observable, cost-aware production systems.
STACK
Skills & tools
Programming
ML & AI
MLOps & Cloud
Data & Backend
Frontend
Languages
WORK
Experience
Jul 2025 — Present
Machine Learning Engineer
Provectus · San Francisco, CA · Remote
Delivering LLM-powered features end-to-end across the stack — Python ML backend, AWS serverless infrastructure, and React/TypeScript frontend — for enterprise document-processing platforms.
- Built production RAG-powered conversational assistants over structured extracted data using LangGraph, AWS Bedrock, and PostgreSQL with pgvector.
- Designed scalable serverless extraction pipelines on AWS Step Functions, Lambda, Textract, and S3, with Alembic-versioned PostgreSQL and full SAM-based infrastructure-as-code.
- Built model evaluation and optimization workflows with MLflow 3.4, DSPy-based LLM-as-judge, embedding-model benchmarking, and human-in-the-loop feedback loops.
- Owned end-to-end delivery across ML backend, serverless infra, and React/TypeScript frontend for document-intensive enterprise pipelines.
PythonAWS LambdaStep FunctionsSAMBedrockSageMakerTextractS3DynamoDBPostgreSQL (pgvector, Alembic)MLflowDSPyLangGraphPydanticReact/TypeScriptJul 2023 — Jun 2025
Growth & Software Development Engineer
JustPaid.ai (YC-W23) · San Francisco, CA · Remote
Part-time agentic AI and full-stack development. Designed and shipped agentic workflows that turn unstructured PDFs into validated, structured records inside the product.
- Designed and implemented agentic AI workflows for automated information extraction from unstructured documents (PDFs), including NLP-based text extraction, structured parsing into Pydantic models, and self-validation loops with agent-based quality supervision.
- Built end-to-end data pipelines integrating ML models with PostgreSQL databases — preprocessing, classification, and structured storage of extracted entities (customers, contracts, line items).
PythonLlamaIndexLangGraphDjangoPostgreSQLReference: Daniel Kivatinos · daniel@kivatinos.com
Aug 2023 — Jun 2025
Teaching Assistant
Universidad de los Andes · Bogotá, Colombia
Led laboratory sessions on machine learning fundamentals: optimization techniques, linear and logistic regression, analytical OLS solutions, hyperparameter tuning, and neural network architectures.
- Managed ~70 students per semester; authored lab resources and assessed student work.
- Average student rating: 4.90 / 5.00.
PythonRscikit-learnTensorFlowPyTorchReference: Luis Felipe Giraldo Trujillo · lf.giraldo404@uniandes.edu.co
BUILT
Featured projects
Research · Multimodal ML
MSc Thesis: Colombian Sign Language Analysis
Recognize, classify, and biomechanically characterize Colombian Sign Language (LSC), and differentiate deaf signers from interpreters using a multi-perspective sensor stack.
Collected multimodal data from deaf signers and LSC interpreters via egocentric vision, conventional cameras, IMU, and EMG in static and conversational settings. Trained Random Forest, KNN, and Gradient Boosting on raw statistical, temporal, and spectral features, and fine-tuned a Video-Visual Transformer (ViViT) on raw video. Reached 95% accuracy on deaf-vs-interpreter classification and 40% accuracy on 50-sign recognition.
Healthcare · Clinical ML
ML for Urological Disease Diagnosis
Identify and explain diagnostic disagreement among urologists at Fundación Santa Fe de Bogotá, and use the model to drive consensus.
Built decision-tree models to flag discrepancies in urologic disease diagnostics across clinicians and extracted feature importances to explain disagreement. Inter-clinician agreement improved from 50% to 75% after model-informed standardization sessions, with feature importance feeding directly into an improved diagnostic workflow.
Web · Open source
Personal Portfolio
Build a fast, accessible single-page portfolio that reflects the work, not the template.
Next.js 15 App Router with Tailwind v4, shadcn/ui primitives, Framer Motion reveals, and a strict dark palette. Deployed on Vercel.
WRITING
Publications
Gomez S, et al. “Machine Learning Analysis of Colombian Sign Language: Recognition, Classification, and Biomechanical Characterization.”
Journal article
Under reviewGomez S, et al. “AI-Based Platform for Automated Uroflowmetry Curve Morphology Classification.”
ICS-EUS 2025
Conference abstractGomez S, et al. “Improving Uroflowmetry Interpretation: Effects of Standardization Sessions on Interobserver Agreement and AI Model Consistency.”
ICS-EUS 2025
Conference abstract
STUDY
Education
Aug 2023 — Dec 2025
MSc in Biomedical Engineering
Universidad de los Andes · Bogotá, Colombia
ThesisMachine Learning Analysis of Colombian Sign Language: Recognition, Classification, and Biomechanical Characterization.
CourseworkMachine Learning for Engineering · Reinforcement Learning · Analysis & Processing of Medical Images
GPA 4.78 / 5.00Jan 2019 — Dec 2022
BSc in Biomedical Engineering — Minor in Neuroscience
Universidad de los Andes · Bogotá, Colombia
CourseworkData Structures & Algorithms · Scientific Programming · Signal Processing · Neuroscience · Neuroanatomy
GPA 4.08 / 5.00
CREDENTIALS
Certificates
AWS Certified Machine Learning Engineer — Associate
Amazon Web Services
Jan 2026
AWS Cloud Practitioner Essentials
Amazon Web Services
Jul 2025
Rapid Application Development with Large Language Models (LLMs)
NVIDIA
May 2025
Efficient Large Language Model (LLM) Customization
NVIDIA
May 2025
Building LLM Applications with Prompt Engineering
NVIDIA
Apr 2025