Kajal Waldiya

- Designed and built NLP workflows for unstructured data, including named entity recognition, keyword extraction, and translation of non-German text - Improved existing GNN-based fraud detection models through targeted feature engineering, resulting in enhanced model performance - Applied LLMs and statistical methods to extract actionable insights from data, supporting investigative analysis of fraud detection - Developed data visualisations to communicate findings to research and stakeholder teams

- Developed features for a core product using Python and Javascript. - Delivered a Proof of Concept for an AI-powered test automation dashboard using reportportal.

- Reduced test script execution time by 65% through multithreaded code optimisation in Python - Automated streaming telemetry data collection for Juniper routers using gRPC and gNMI protocols - Built an automation tool using Python and Robot Framework

- Delivered end-to-end development and deployment of an IT helpdesk chatbot using LUIS for an ASEAN client - Built a PoC for invoice processing using deep learning–based OCR - Developed CMS and REST APIs using Django for a media company - Applied image preprocessing, OCR, and NLP techniques to production systems adopted by internal client users

- Developed web analytics tracking and server monitoring platform using Python - Worked on text extraction from digital IDs for on-demand ID card generation - Data visualization for gathered user behaviour and extract insights - Implemented Face detection and extraction from digital identities for identity verification - Developed application for machine learning products