Leuven, Flemish Region, Belgium
Head of AI + Builder Love building fun stuff, stuff of the future. Master's in AI, KU Leuven. Bachelor's in CS, IIT Hyderabad
Leading AI and data product teams in the Belgian tax and accounting space. Setting company-wide AI strategy and building internal tooling that makes everyone's work easier. What my team and I are building: Document Processing: GenAI, traditional ML, and yes, even regexes. Processing and extracting data from mission-critical tax documents at 98%+ accuracy across 18 document types. Web Scraping: Retrieving mission-critical tax documents and data (100x speed improvements underway). KYC & Compliance: Automated AML systems combining deep research with web scraping. Cross-functional Collaboration: Working closely with CTO, CEO, Product, and Sales to embed AI across the business. Internal Tooling: Service desk ticket analysis for faster UI/UX feedback loops, support article suggestions, and BI dashboards, sales data analysis, etc. AI-Assisted Coding: Driving adoption strategies across engineering.
Currently busy integrating LLMs into our Invoice Recognition software. In my role at Ixor, I engage in a variety of tasks within AI and software development: AI Development: Focused on enhancing AI functionalities, including the integration of GPT with Retrieval Augmented Generation, and working on Convolutional Neural Networks (CNNs) and Natural Language Processing (NLP) for optimizing our invoice recognition software. These efforts have significantly improved the software's processing capabilities and intelligence. Backend Development: Contributed as a core team member to the backend development of projects such as Udini, IxorDocs, and the Belgium Traffic Center’s monitoring software, emphasizing scalable and efficient software solutions. Full Stack Contribution: Extended my role to include frontend development, ensuring cohesive and user-friendly software interfaces. Technological Versatility: Employed a diverse set of technologies, including Python (PyTorch), Java (Spring Boot), JavaScript/TypeScript (React), and AWS (ECS, Lambda, ECR), adapting to various project needs. Team Collaboration: Worked in tandem with interdisciplinary teams, merging AI research with practical software development to create high-quality software products. My tenure at Ixor has been characterized by continuous learning and application of AI technologies in dynamic and innovative ways.
Project: Explainability of deep learning models Responsibilities: Making the results of deep learning models more understandable using heatmaps and trust scores Technologies used: Keras, Tensorflow, OpenCV Mentor: Elen Bart, VITO
Project: Building an end-to-end causal analytics framework Responsibilities: Research and develop a python library for solving potential marketing problems of causal nature Skills acquired: In-Depth understanding of causality and its often overlooked repercussions in the industry Technologies used: Python, scikit-learn Mentor: Mr. Pranav Verma, CEO, Busigence Technologies
Project: Building Machine Learning pipelines to detect Dry Eye illness Responsibilities: Developed the software end-to-end, from image processing and feature extraction on raw image data to training ensemble ML models Skills acquired: Building ML pipelines in Python Technologies used: Python, MATLAB, sklearn, Tensorflow Mentor: Mr. Kumar Rajamani, BUD Dept, Robert Bosch Engineering