Senior Computer Vision Engineer | Machine Learning | Robotics
7+ years building and deploying end-to-end computer vision and deep learning systems — from research prototypes to production pipelines serving hundreds of clients. I've shipped real-time perception systems at startups (Arvist, Aren.AI), led CV teams at Deloitte, and built vision solutions for a major smartphone manufacturer at Eldorado Research Institute.
My core loop: train a model, optimize it (TensorRT, ONNX, OpenVINO), deploy it to edge or cloud, and wire up the MLOps so it keeps improving without manual intervention. I also maintain CondadosAI, where I share dev tools, utilities, and tutorials for computer vision, AI, and robotics.
Intel AI Innovator · B.Sc. Computer Engineering, Summa Cum Laude (UFRN) · Best Computer Engineer Graduate 2020
| Domain | Technologies |
|---|---|
| Deep Learning | PyTorch, TensorFlow, scikit-learn, self-supervised learning |
| Computer Vision | OpenCV, Open3D, 3D reconstruction, point clouds, object detection (YOLO), segmentation |
| Model Optimization | TensorRT, OpenVINO, ONNX, OpenMP |
| ML Frameworks | MMDetection, MMSegmentation, MMDeploy |
| MLOps & Deployment | MLFlow, Prefect, Docker, FastAPI, edge deployment, active learning pipelines |
| Robotics & Control | Path planning, PID control, Kalman filters, RTOS, CoppeliaSim, CARLA |
| Languages | Python, C, C++ |
| Cloud & Infra | AWS, GCP, Azure, Linux, Git, Bash, Grafana |
| Role | Company | Period |
|---|---|---|
| Senior Computer Vision Engineer | Arvist (Techstars '22) | 2025 -- present |
| Senior Computer Vision Engineer | Aren.AI | 2023 -- 2025 |
| CV Engineer · Senior Consultant IV | Deloitte | 2023 |
| CV Researcher / Engineer | Eldorado Research Institute | 2021 -- 2023 |
| CV & Robotics Researcher | Robotics Lab, UFRN | 2018 -- 2020 |
Complete software stack for a competitive robot soccer team — real-time vision, path planning, motion control, and wireless communication, all running together in the loop. Built with the Poti team at UFRN. C C++ Qt V4L2 Bluetooth Path Planning
Perception-to-control autonomous driving pipeline developed as the capstone for the Coursera Self-Driving Car Specialization. Python CARLA Simulator Autonomous Driving
My undergraduate thesis. Designed a feedforward/feedback control scheme for robots with low-resolution magnetic encoders — Kalman filter for state estimation, PID for actuation, all running on an ESP32 under RTOS. C C++ ESP32 RTOS Kalman Filter PID
Ready-to-deploy MLFlow setup for experiment tracking and model versioning. Python MLFlow Docker MLOps
Robotics & Simulation Projects
- Path Planning with Potential Fields — C++ potential-field planner running in CoppeliaSim, written from scratch. Source
- Probabilistic Occupancy Grid — Grid mapping algorithm implemented from scratch in C++. Source
- Kinematic Controllers — Path following, position control, and trajectory tracking in C++
- Configuration Space Collision Detection — Custom collision checker in configuration space, implemented from scratch
- Robot Firmware — IEEE VSSS — RTOS-based ESP32 firmware for the Poti competition robots
Computer Vision & Image Processing
- OpenCV & Python Projects — Collection of hands-on projects: QR reader, face detection & recognition, YOLO object detection, OCR, facial landmarks, and more
- Digital Image Processing Portfolio — Coursework projects covering histogram equalization, motion detection, tilt-shift, K-means quantization, and color-based tracking
- 2D Convolution from Scratch — C++ convolution operator with configurable stride and padding, built on a custom PGM image class
- Adaptive Blur Filter — Content-aware blur that adapts kernel size based on local image features
- Hough Transform for Squares — Geometric square detection via Hough transform in C++
- Art with Threads — Generates thread art (Art Nitka style) from any input image through exhaustive line-placement search
Model Deployment & MLOps
- OpenVINO Exploration — Benchmarking and optimizing models with Intel's OpenVINO toolkit
- TensorRT Exploration — GPU inference acceleration experiments with NVIDIA TensorRT
Graphics & Other
- Ray Tracer from Scratch — Software renderer in C/C++ with OpenMP for parallel ray casting
- Bluetooth C/C++ Library — Lightweight library for Bluetooth communication on Linux
- Simplex Method — Python implementation of the simplex algorithm for linear programming
- Transport Problem & Nonlinear Analysis — Simplex extensions applied to transport problems and nonlinear function analysis