AI/ML Engineer · Engineering Physics graduate · Campinas, São Paulo, Brazil
I build production-oriented AI systems across documents, business workflows, backend services, cloud environments, and connected devices. My background spans RAG, multi-step agents, Python APIs, PostgreSQL, embedded firmware, and real-time machine learning.
- Building AI-assisted document and workflow automation with Python, TypeScript/React, PostgreSQL/Supabase, APIs, and human-review controls.
- Developing retrieval and agent-integration components, evaluation and regression workflows, automated document generation, Docker tooling, and runtime reliability improvements.
- Exploring reliable agentic systems that combine LLM reasoning with deterministic tools, structured evidence, validation, and auditable delivery.
- Agentic AI: Implemented multi-step document workflows covering context acquisition, task routing, generation, validation, approval, versioning, and delivery.
- RAG and LLM systems: Built vector-retrieval workflows with FAISS, PostgreSQL/pgvector, Qdrant, embeddings, semantic search, prompt evaluation, and evidence-grounded responses.
- Edge ML: Led the machine-learning component of a real-time energy load-disaggregation system operating at 10 Hz with sub-two-second inference latency.
- Embedded systems: Developed STM32/ARM Cortex-M4 firmware with FreeRTOS, interrupts, timers, sensor interfaces, telemetry, OTA support, and watchdog/recovery behavior.
- Backend and cloud: Worked with FastAPI, Flask, Django, Supabase, AWS, Azure-hosted LLM services, Docker, REST APIs, CI/CD, and automated testing.
- multilingual-multihop-rag — Local-first multilingual RAG using FastAPI, pgvector, evaluation tooling, and local LLMs.
- Robotica_dacardenasj — STM32/FreeRTOS mobile robotics with IMU heading control, PID, odometry, telemetry, and on-device A* path planning.
- STM32CortexM4ElectronicaDigital — Bare-metal ARM Cortex-M4 drivers, peripheral interfaces, displays, and memory-constrained applications.
- Transcribe-assemblyai — Chunked audio transcription with overlap handling and a responsive web interface.
- personal-site — Next.js and MDX portfolio documenting selected AI, backend, embedded, and academic projects.
- Languages:
Python·C/C++·TypeScript/JavaScript·SQL·Go - AI and data:
RAG·LLM agents·FAISS·PostgreSQL/pgvector·Qdrant·Pandas - Backend and cloud:
FastAPI·Flask·Django·React·Supabase·AWS·Azure·Docker - Systems:
STM32·ARM Cortex-M4·FreeRTOS·ESP32·MQTT
- Ingeniero Físico (Engineering Physics) — Universidad Nacional de Colombia, 2025; cumulative GPA 4.5/5.0.
- Former Teaching Assistant in mobile robotics, ARM microcontrollers, embedded systems, simulations, and mobile technologies.
- IELTS Academic — Overall Band 7.0, CEFR C1, July 2026.
- Spanish native; English C1.