Agentic Chatbot Application
Architecting and leading the development of an AI-first solar advisory and onboarding platform to streamline customer journeys and internal workflows. The first application created is a chatbot which not only generate Leads but also helps user in getting answers for anything related to Solar. This application is integrated with calculator, CRM application for automated processes. I have designed and implemented agentic workflows using OpenAI and Gemini models, automating user query handling, and multiple system navigation. This application reduce time-to-delivery by 60%. Application is created using OCR, Python, MongoDB, and custom APIs. This application reduces manual intervention by 70% through LLM-driven task orchestration and workflow automation.
Electronic Program Guide & TV Search
We aggregated and enriched TV program schedules from across India, delivering high-quality metadata to broadcasters, cable operators, and online platforms. Our solution ensured accurate, real-time updates, enhancing content discoverability and viewer engagement. The data was provided in XML, JSON, Excel, and TXT formats, ensuring seamless integration with client systems. With a flexible entitlement-based model, we supplied 7, 14, or 30 days of scheduled data, catering to diverse business needs while ensuring data security, compliance, and operational efficiency.
Apps, Portals & Recommendations
We developed a proprietary Recommendation Engine, deployed on our portal and mobile apps, and offered it as a service to broadcasters and cable operators. This platform integration enhanced content discovery, allowing users to easily find programs and channels while enabling clients to upsell premium packages and channels. By leveraging personalized recommendations, we helped increase ARPU (Average Revenue Per User) by 15-25%, driving higher customer engagement and revenue growth. We used ASP.net, C#, Java and SQL server as our technology stack.
TV Street Maps
We developed a proprietary Recommendation Engine, deployed on our portal and mobile apps, and offered it as a service to broadcasters and cable operators. This platform integration enhanced content discovery, allowing users to easily find programs and channels while enabling clients to upsell premium packages and channels. By leveraging personalized recommendations, we helped increase ARPU (Average Revenue Per User) by 15-25%, driving higher customer engagement and revenue growth. We used ASP.net, C#, Java and SQL server as our technology stack.
The Unified Data Ingestion System (UDIS) was envisioned by me in 2016 and architected from scratch to centralize all ingestion processes across Gracenote’s verticals—Video, Music, and Sports. I built and led the core team to develop this system with a strong business use case, ensuring scalability, efficiency, and future-proofing.
Key Highlights:
✅ Initial Development (2016): Built using PHP, MySQL, and S3 technology.
✅ Current Architecture: Running on AWS with Airflow as the orchestration layer, and all services deployed as microservices on Kubernetes for scalability, resilience, and automation.
✅ AI-Powered Enhancements: Recently integrated Claude + LLM to automate data mapping, reducing manual intervention by 90%.
Since its inception, UDIS has transformed Gracenote’s ingestion processes, providing a unified, scalable, and intelligent data pipeline that enhances operational efficiency and content management across all business verticals.
The Electronic Programming Guide (EPG) system, acquired from a European provider, was expanded and optimized by my team to support global ingestion across all Gracenote-managed regions, including APAC, Europe, U.S., LATAM, and Halifax. This standalone, end-to-end application not only handled ingestion but also monitored and managed data requests from 1,500+ editors at Gracenote. It provided real-time insights into source failures, content modifications, and format changes, ensuring seamless metadata processing, enhanced operational efficiency, and improved content accuracy across all markets.
Relay is a legacy Java-based application inherited through the Tribune acquisition, primarily managing Linear channels across the U.S. and LATAM along with VOD (Video on Demand) services. Post-acquisition, my team and I enhanced and optimized the application by removing outdated Ruby-based ingestion code and migrating all Linear channels to the EPGs platform. We further expanded Relay’s capabilities to handle more ingestion sources, including:
✅ Linear channels globally via EPGs to Relay
✅ VOD ingestion directly via Relay
✅ OVD (Over-the-Top Video Distribution) via UDIS → EPGs → Relay
These enhancements streamlined data ingestion, improved system efficiency, and ensured seamless content management across multiple platforms.
We built a Content Management System (CMS) from scratch, designed to streamline the end-to-end content workflow—from ingestion to enrichment and delivery. The system efficiently processed raw data files from multiple sources, storing them in a normalized database to maintain master records for program names, metadata, languages, and channels. Our content editors enriched the metadata by adding missing information, correcting literature, and incorporating imagery. This unified platform served as a single source of truth, ensuring data accuracy, consistency, and seamless distribution to broadcasters, cable operators, and digital platforms.
CitiStreet was a joint venture between Citigroup and State Street, offering asset management, outsourcing, and administrative services for pension, health, and welfare plans. I worked with this client from Bangalore and Jacksonville, U.S., leading project development in .NET technologies while also supporting and modernizing legacy applications.
Key Contributions:
✅ Led development and enhancement of core applications in ASP.NET, C#, and SQL Server.
✅ Migrated and re-engineered multiple insurance-related software from MS Access to modern .NET architecture, improving performance and scalability.
✅ Provided end-to-end development and technical leadership, ensuring seamless integration and minimal downtime during migration.
This project streamlined CitiStreet’s operational processes, enhanced system efficiency, and modernized critical financial applications for better maintainability and scalability.
I worked with Avaya to develop automated customer care solutions for their telephony services, assisting in both solution development and pre-sales activities.
Key Contributions:
✅ Designed and developed a robust Java-based solution with extensive permutation and combination logic to handle diverse customer interactions.
✅ Enabled automation for customer support across multiple touchpoints, improving response efficiency.
✅ Extended the solution for use in various IP-based telephony services, enhancing Avaya’s service offerings.
This solution streamlined Avaya’s customer service operations, reduced manual intervention, and enhanced the efficiency of their IP-based telephony ecosystem
UBS – Union Bank of Switzerland was my first overseas project in Singapore, where I worked on Unix Shell Scripting, PHP scripting, and SQL Server.
Key Contributions:
✅ Designed and implemented a robust backup system to ensure rapid and secure data replication across UBS's four major hubs.
✅ Optimized database synchronization across multiple locations, ensuring minimal downtime and data consistency.
✅ Developed and enhanced automation scripts to improve the efficiency and reliability of backup processes.
✅ Ensured data availability for high-net-worth clients, where real-time access and failover readiness were crucial.
This project played a critical role in UBS’s global data management strategy, improving disaster recovery capabilities and ensuring seamless financial operations across regions.
This was my first major project, developed and delivered by a 45-member team. I contributed as a Software Developer, implementing solutions using Java with Oracle as the database.
Key Contributions:
✅ Developed critical modules for automating the supply chain pipeline of a historic European shipping company.
✅ Collaborated with a large team, ensuring seamless design, development, and integration.
✅ Optimized database queries and implemented scalable backend logic to enhance system performance.
✅ Project Duration: Successfully delivered in 10-12 months.
✅ Recognition: Received the "Best Project Execution Award" at L&T Infotech for outstanding delivery and execution.
This project enhanced operational efficiency, automated logistics and inventory management, and streamlined the client’s supply chain operations.