Leading the firm's global AI/ML function, owning strategy, talent, and platform execution across a distributed international team. Defined and executing a multi-year enterprise AI roadmap, establishing a scalable, multi-tenant Generative AI platform as the firm's core infrastructure for RAG and Agentic AI workloads.
- Championed adoption of frontier AI capabilities — multimodal models and domain-adapted solutions across financial research, risk, and operations
- Drove executive-level governance through WBR/MBR cadences, establishing platform metrics, adoption KPIs, and goals reported to C-suite
- Partnered across investment, technology, and business functions to embed AI as a firm-wide strategic capability
- Global org design across AI/ML research, engineering, and platform functions
Amazon Bedrock GenAI
Two landmark roles across nearly 8 years at AWS. As Head of Engineering for GenAI Applications on the Amazon Bedrock team, built the end-to-end stack for Agents, Flows, Guardrails, Evaluation, Prompt Management, and Automated Prompt Engineering — leading 100+ people. Prior to that, as ML Head for Amazon Lex, built NLU and ASR systems from scratch and grew the team from 7 to 100+.
- Built Bedrock Agents and Guardrails — groundbreaking services that set new industry standards for agentic AI and safety
- Led a scalable platform to evaluate foundational models across dimensions, sizes, and modalities
- Built 30+ microservices for Amazon Lex, patented 20+ ideas, and launched deep learning models at scale
- Took Lex global — launched in 10+ regions and 25+ locales across 24 months
- Co-owned the Lex P&L, drove 100% yearly revenue growth for 3 consecutive years, and brought the service to profitability
- Reduced latency and costs by up to 70% through architectural simplification
Part of the Getaways team, managed the Inventory Platform responsible for inventory, availability, pricing, and reservations for all travel deals across Groupon worldwide. Built the platform from scratch to integrate with multiple channel managers and third-party wholesalers across the US and EMEA.
- Grew the team from 4 to 20 people, attracting top talent internally and from the industry
- Created a 5-year vision and multi-year roadmap; drove long-term and short-term business strategy
- Built the supply platform from the ground up integrating with multiple partners worldwide
- Unified platforms to reduce redundancy and improve reliability, maintainability, and resiliency
- Promoted top talent and mentored several engineers and managers
Part of the Amazon Prime and Vendor Payments group. Owned Delivery Messaging for Amazon customers across all pages on amazon.com — serving all locales and all devices worldwide.
- Built a low-latency, scalable service for all delivery messaging features across amazon.com worldwide
- Centralized Delivery Messaging across pages without regressing latency on the high-traffic detail page
- Built a framework for messaging, content experimentation, and A/B testing
- Migrated all features to responsive web — PC, mobile, and tablet
- Cleared technical debt and built scalable, long-term solutions while driving down latency
Worked on Windows 8 Server in the networking team, part of the manageability team providing a uniform and easy way to manage DNS components on Windows Server.
- Developed PowerShell cmdlets to manage DNS activities
- Developed DNSSEC UI for Windows 8 DNS Server
- Ensured IPv4 and IPv6 compatibility across all cmdlets
Worked on publishing platforms to ensure Yahoo! had the best publishing framework used across various applications — serving the right content to people at the right time.
- Built a generic Role-Based Access Control system for content management
- Developed a content scheduling system used across Yahoo! publishing properties
- Created a seamless release process for all deployments
- Built platforms enabling website creation in a few clicks
Public projects on GitHub — from agentic AI frameworks to hands-on demos for understanding how modern AI systems work.
Patent applications across AI, ML, and cloud technologies
Filed primarily during the Amazon Lex and Bedrock years — inventions spanning natural language understanding, automated speech recognition, agentic AI systems, and cloud infrastructure. A track record of identifying the ideas worth protecting at the frontier of enterprise AI.
Innovations in NLU, intent recognition, slot filling, and conversational AI systems at enterprise scale.
ASR systems, speech-to-text pipelines, and real-time audio processing architectures built for Amazon Lex.
Agentic AI frameworks, prompt optimization, model evaluation systems, and guardrail architectures from the Bedrock era.