Tomás Echavarría Bayter Email & Phone Number

Colombia

Medellín, Antioquia, Colombia

- Developed and implemented instance segmentation models (Detectron2, YOLO) to identify more than 300 food categories.- Implemented microservices and APIs (FastAPI) for hosting diverse AI models (NLP and CV), ensuring efficient deployment and seamless integration.- Proof of Concept (POC) for AI implementations using Gradio interfaces.- Leveraged state-of-the-art LLMs for innovative NLP services, including content generation, structured/unstructured database search, audio processing, and NLP model deployment.- Designed and implemented a chatbot for customer support inquiries using Retrieval Augmented Generation (RAG).- Performed fine-tuning and optimization on Stable Diffusion, DALL-E 2, and various other generativealgorithms to enhance the creation of high-fidelity synthetic images.- Established a robust CI/CD pipeline for seamless model training and deployment across different environments.- ETL process for handling unstructured data, enabling efficient data analytics and visualization.- Migrated services from AWS cloud to GCP.

Medellín, Antioquia, Colombia

- Focused on providing consulting services in analytics and data science to SMEs in different sectors tosupport strategic decision-making.- Main projects developed: · AI model for predicting aggression in children through graphology (PyTorch). · Psychological testing evaluation software. · Insights generation for data from companies in the Bogot´a Chamber of Commerce. · Consulting for a real estate company.

Medellín, Antioquia, Colombia

- Developed a composite AI model for currency price and trading signal prediction (TensorFlow).- Researched and analyzed historical data from different currencies.- Generated trading indicators and studied trading methodologies.- Conducted testing, validation, and model refinement.

Medellín, Antioquia, Colombia

- Developed Artificial Intelligence (AI) algorithms and data analytics for the project ”Web platform for datacollection, visualization, analysis, prediction, and evaluation of strategies for controlling the disease caused by SARS-CoV-2”.- Designed and implemented an ETL process for Covid-19 historical data.- Developed and deployed a predictive model to assess the mortality risk associated with Covid-19.

Medellín, Antioquia, Colombia

- Led the expansion and advancement of telemetry and IoT projects.- Processed and analyzed data obtained through mobile devices (Python).- Established communication channels through APIs and databases to integrate with a visualization platform.- Designed and deployed remotely connected devices.

Envigado, Antioquia, Colombia

- Developed an electronic characterization system for piezoelectric crystals used in microbalances for analyte identification in biological samples.- Created and implemented a processing software, including a graphical interface, data processing, andserver-controlled system.

Rochester, Minnesota, Estados Unidos

- Conducted a research internship in the Department of Physiology and Biomedical Engineering.- Created a mathematical model to evaluate the feasibility of developing a technique for measuring a new mechanical parameter for ultrasound imaging.