Mar 2022 – Jun 2024 · 2 yrs 3 mos · Bengaluru, Karnataka, India

• Spearheading a team of AI experts in developing a cutting-edge Generative AI Software Developer Co-Pilot, enhancing coding efficiency with proprietary GenAI algorithms.
• Architecting an innovative Gen AI Co-Pilot, adept at generating automated unit tests, leveraging reinforcement learning, and offering multi-language support for streamlined DevOps integration.
• Pioneering self-improving solutions through self-learning algorithms complemented by user feedback mechanisms.
• Crafting and executing a strategic product vision and roadmap in alignment with organizational objectives and market dynamics.
• Fostering collaboration with cross-functional teams to guarantee the successful launch and market penetration of AI products.
• Serving as a Generative AI thought leader, advising on emerging technologies and best practices within the industry.
• Supervising the creation of advanced Generative AI algorithms to ensure peak performance and scalability.
• Contributing to the creation of a zero-code platform for constructing audio/video media pipelines compatible with various Intel processors and supercomputing environments.
• Implementing best-in-class AI/MLOps practices and tools to secure the longevity and success of data science initiatives.
• Leading the charge in MLOps integration, automating model deployment on Azure Kubernetes, and establishing secure CI/CD pipelines.
• Leading the development of the GStreamer plugin to perform inference of real-time data analytics, including video analytics at scale and store the inference results in a scalable database for debugging models supporting Intel's in-house deep learning framework- OpenVino optimised for CPU

May 2019 – Mar 2022 · 2 yrs 10 mos · Bengaluru Area, India

• Led a robust team of AI experts and software development, driving AI innovation and infrastructure optimization to deliver custom AI solutions for complex business needs.
• Developed Complex in-house AI models utilizing deep learning techniques for artistic style transfer, precise font recognition, and efficient QR code detection and decoding.
• Advanced text analytics with zero-shot classification using BART, AWS text enhancement, and multilingual Named Entity Recognition (NER) development.
• Streamlined AI workflows on AWS, achieving cost-effective model deployment and playing a pivotal role in technical talent acquisition.
• Orchestrated the development of scalable pipelines for AI model hosting on AWS, incorporating both online and offline training, and transitioned to ECS and EKS for improved autoscaling capabilities.
• Engineered robust infrastructure to support large-scale AI models on multi-GPU environments and actively participated in the hiring process for data science and software positions.

Apr 2018 – May 2019 · 1 yr 1 mo

• Spearheaded the creation of an AI-driven Vehicle Route Optimization Algorithm employing multi-constrained discrete optimization and MILP techniques, tailored to adhere to complex business requirements.
• Engineered a Truck Loading Plan algorithm, leveraging unsupervised machine learning to efficiently manage heavy cargo within optimized routes, despite limited classification data.
• Enhanced route optimization realism by incorporating GPS data from fleet vehicles, utilizing DBSCAN clustering and MILP for expedited processing.
• Formulated a Warehouse Management Algorithm to optimize spatial efficiency, applying diverse learning methodologies to solve logistical challenges.
• Currently advancing an algorithm that integrates GPS data with various ML models to further refine route optimization strategies.
• Crafted AI vehicle routing algorithm, enhancing path efficiency with MILP and business constraints.
• Integrated GPS data with ML for realistic route optimization, reducing time with DBSCAN and MILP.
• Devised warehouse management algorithm to boost space use, employing varied learning techniques.

Feb 2015 – Apr 2018 · 3 yrs 2 mos · Bengaluru Area, India

• Spearheaded the development of Reverse Auction Engine for an agriculture-based online marketplace, leveraging Random Forest Ensemble models to predict bids and optimize farmer selection based on capacity and SKU offerings.
• Engineered and benchmarked multiple classification algorithms, including Logistic Regression, Decision Trees, K-Nearest Neighbors, and Support Vector Machines, to enhance predictive accuracy.
• Implemented K-Means clustering to detect outliers and categorize unlabeled data, improving data integrity and model performance.
• Analyzed farmers' bidding patterns using statistical modeling to refine the auction engine's selection process and maximize participation efficiency.
• Created a Machine Learning-driven Recommendation Engine to match buyers with suitable produce from various food hubs, enhancing customer satisfaction and sales.
• Optimized vehicle routing using A* search algorithm tailored to vehicle size, multi-point pickups, and delivery locations, resulting in cost and time savings.
• Designed an efficient palletization system to accommodate multiple buyers and SKUs, aligning pallet sequences with optimized truck routes for streamlined logistics.
• Managed the USDA CropScape Data web application, processing satellite imagery data for 200 crops with 8 million data points annually, utilizing Python libraries and Hadoop ecosystem tools for scalable data handling.
• Forecasted annual crop production across the United States at a granular level using ARIMA and FB Prophet, leveraging a decade of USDA historical data to inform agricultural planning.
• Developed a web scraping solution to extract USDA Price Lists from 100 HTML pages daily, employing Beautiful Soup and Selenium in a serverless architecture, with data storage managed in MongoDB.

Jan 2012 – Jan 2014 · 2 yrs