- Clean and preprocess geospatial data obtained from various sources, ensuring data quality and integrity.- Integrate and transform diverse datasets, including soil data, into a unified geospatial format suitable for analysis.- Utilize ArcGIS Pro to analyze geospatial data and identify patterns, trends, and correlations related to valley fever occurrence

- Led the infrastructure development of a patent search web app, managing 4 million documents and implementing features for monitoring, analyzing, and predicting market trends - Fostered agile collaboration across interdisciplinary teams, leading stakeholder discussions and refining project requirements. Ensured successful end-to-end execution for project delivery - Applied design patterns and SOLID principles to develop over 20 REST API services for text search, document retrieval, and data visualization using Python FastAPI - Fine-tuned transformers model and coded 10+ NLP modules such as semantic search, topic clustering, keyword 
extraction, and text classification using custom RNN’s and Word2vec - Established a microservices core employing OAuth for user authentication, Elasticsearch for accelerated indexing, and cloud hosted ML model for query search - Enhanced the user experience of a frontend application by skillfully utilizing Vue.js, D3.js, and Highcharts, ensuring adherence to the MVC architectural pattern - Part of the MLOps prototype team with goal to Automate ML training workflows, CI/CD for automatic 
integration and deployment using MLflow, kubernetes and Seldon- Transformed a struggling legacy data pipeline into a microservices-based architecture, resulting in a significant 40% 
reduction in runtime. Implemented database Views to enhance query performance and streamline data retrieval - Applied transfer learning on MobileNetV2 and fine-tuned it to classify obstacles in a lidar point cloud image with 
98% accuracy. Effectively used mlFlow to track model training experiments over time

- Delivered an AI-based analytics tool atop website content of 1.5 million companies. This tool was adopted by the Strategy team at Infineon to identify organizations with overlapping hardware and software capabilities - Utilized serverless computing on AWS to deploy a weekly batch scheduler. It extracted, processed, and enriched data from websites, storing the results in an Amazon S3 bucket and NoSQL database for analysis - Created multiple data visualization pages and views utilizing D3.js, showcasing data that were dynamically 
updated in real-time, ensuring continuous access to current and relevant insights - Transitioned a monolithic chatbot codebase to a modular framework with asynchronous programming, achieving a 
30% improvement in response time - Collaborated across departments to deploy more than 5 tailored chatbots, resulting in a 20% increase in the productivity of design engineers

Projects : MTA : Web-app to assist the Strategy team at Infineon in searching and localizing trends related to specific keywords in the semiconductor industry. -Designed a web-app using using python flask as backend. Used html, JavaScript and bootstrap for the GUI- Enhanced the GUI with interactive graphs from D3.js library showing sentiments , document distribution across region and industry - Created a scheduler running on AWS to extract the articles weekly, dump them into MySQL usingbeautifulsoup and selenium- Followed Test Driven Development/Behavioral Driven Development while writing unit tests in pyUnit- Deployed the tool on OpenShift Cloud platform for access across the organization

- Implemented an AI search tool to identify best fit companies for synergy across product and application lines- Extracted and aggregated data for organisation from third party REST API’s- Used REST API’s to extract, process and store market intelligence data from third party data aggregators . e.g. Crunchbase , Capital IQ- Trained multiple Doc2Vec models and devised a method for UAT (User Acceptance Test) to evaluate the model performance