.png)
I'm exploring how we can make agents see, understand, and transform the world around us. I’m dedicated to building systems that bridge perception and action, shaping a future where technology seamlessly integrates into our lives.
Main Innovations:
1️⃣ Comparative analysis of RGB vs. thermal imagery.
2️⃣ Auto-labeling to simplify data preparation.
3️⃣ Real-world applicability of the framework.
.png)
1️⃣ Led the design and development of AI-powered software that integrates with security cameras to analyze customer behavior in retail settings, providing insights such as demographic data (gender, age) and in-store movement patterns.
2️⃣ Scaled the solution to monitor over 500 million customer interactions across malls in Colombia, providing demographic and movement pattern analytics.
3️⃣ Built a full production pipeline using TensorFlow, PyTorch, Docker, and Google Cloud, ensuring performance and scalability for seamless deployment.
Tech Stack: OpenCV, RT-DETR, YoloR, OpenCV, TensorFlow, Pytorch, Google Cloud Platform, Docker.
AccuNeRF leverages Neural Radiance Fields (NeRF) technology to transform 2D images into 3D digital models for automotive collision reconstruction. By integrating 'Nerfacto' from Nerfstudio and refining raw data pre- and post-processing, the approach delivers highly accurate representations of vehicular accidents, offering a new standard for real-world scene fidelity in accident analysis.
KayEcho.AI is a full-stack web app leveraging LangChain, Anthropic’s Claude API, and MongoDB Atlas Vector Search to transform professional networking through advanced Retrieval-Augmented Generation (RAG) pipelines. By combining profile scraping, iterative prompt engineering, and simulated conversations, it ensures precise matches and actionable insights for event attendees.
1️⃣ Improved object detection accuracy from 80% to over 94% and expanded the dataset from 0 to 50,000+ images, leveraging TensorFlow, Open3D, and PointNet.
2️⃣ Developed a full production pipeline, deploying real-time object detection and classification models using MobileNetV2, ResNet, and AWS Lambda.
3️⃣ Automated the feedback loop with image embedding (Base64) for labeling validation, enabling seamless user feedback and continuous model improvement.
Oysters are a vital keystone species in coastal ecosystems, providing significant economic, environmental, and cultural benefits. As the importance of oysters grows, so does the relevance of autonomous systems for their detection and monitoring. However, current monitoring strategies often rely on destructive methods. While manual identification of oysters from video footage is non-destructive, it is time-consuming, requires expert input, and is further complicated by the challenges of the underwater environment.
System for detecting anomalies in underwater pipelines, using Deep Learning. Innovate 2019 Winners | Ecopetrol.
Main tasks:
● Create an object detection model to detect objects in manholes.
● Create classification models to classify multiple materials and shape in manholes.
● Create Segmentation models in Point Cloud to classify parts of manholes.
● Create preprocessing algorithms for videos of manholes and pipes.
Tech Stack: TensorFlow, Open3d, PointNet, segmentation models.
Click Subterra information
Main tasks:
● Design agents for races and puzzles games (Reinforcement Learning).
● Experiment with computer vision algorithms to improve gaming experiences.
● Apply AR environment to teaching new languages.
● Created an NLP model that can classify feelings and topics in a real-time chat of games.
Tech Stack: Albert, Roberta, TensorFlow, Reinforcement Learning, AR, MediaPipe, Python.
Main Tasks:
● Led the design and development of AI-powered software that integrates with security cameras to analyze customer behavior in retail settings, providing insights such as demographic data (gender, age) and in-store movement patterns.
Tech Stack: OpenCV, RT-DETR, YoloR, OpenCV, TensorFlow, Pytorch, Google Cloud Platform, Docker
Click Switch Infromation
You already know how neural networks work. Increase your skills using TensorFlow and its entire ecosystem of tools. Create deep learning models that you can put to work in professional environments.
Building and optimizing tools for object detection, image segmentation, and video analytics to advance computer vision research.
1️⃣ Utils to supervision
2️⃣ No Maximum Suppression
3️⃣ ByteTracker in Supervision
4️⃣ Integrate yolo-world and yolo-nas to Fiftyone
Creating tutorials, guides, and educational content to empower developers and students in mastering AI and computer vision.
4️⃣Keras code example 3.0
Are you passionate about innovation and creating impactful solutions? Whether it's AI-powered applications, computer vision breakthroughs, or open-source collaborations, I’m always excited to explore new opportunities.
Ready to work with me?
