AI Engineer
Education
Computer Science & Data Science @ Ateneo de Manila University (2018-2022)
Work Experience
AI Engineer @ IBM (Apr 2023 - Present)
- Generative AI: Designing and developing LLM-powered proof-of-concepts via Watsonx.ai, driving up to 50% productivity gains and a 10% increase in sales profits through innovative Generative AI applications.
- Large Language Model Optimization: Enhancing LLM performance with advanced techniques like prompt engineering, caching, and algorithmic refinements, resulting in measurable improvements in accuracy and relevance.
- Backend Development: Engineering scalable and reliable backend APIs using FastAPI and Flask to support advanced LLM applications.
- Model Deployment: Deploying customized LLM solutions on IBM Cloud via Code Engine, tailoring implementations to meet diverse client requirements and ensuring optimal performance.
Junior Data Scientist @ Kumu (Aug 2022 - Nov 2022)
- Exploratory Data Analysis: Collected data from distributed data processing via Spark, and analyzed data by applying PCA, and pre-processing techniques such as checking for multicollinearity/segmentation - promoting streamer discovery using second-degree connections.
- Model Development: Built a recommendation engine using collaborative filtering and deployed the model on Kubernetes.
- Experimentation: Applied A/B test framework to 100k users.
Projects
LLM-Driven Sales Analytics App
- Developed an LLM-powered PoC via Watsonx.ai, generating data-driven sales insights and a 10% sales profit increase, proving its potential to significantly reduce analysis time for sales leaders by up to 50%.
Legal Contracts Analytics App
- Implemented a legal contracts analytics app via LLM to summarize legal documents, identify potential risks, and generate recommendations for improvement - resulting to 80% faster legal planning and analysis.
RAG-based Credit Risk Assessment System and Governance
- Built a RAG-based credit risk assessment system that predicts borrower default while ensuring model governance for reliable decision-making, reducing credit losses by up to 25%.