Portfolio
Featured Projects
Showcasing work in AI/ML, platform architecture, and scalable systems that create real business impact.
Challenge
New Relic needed an enterprise-grade knowledge platform that could ingest and reason over millions of internal artifacts — repositories, runbooks, documents, and tribal knowledge — while meeting the latency and trust bar for production observability.
Solution
Architecting a large-scale knowledge platform processing data across ~3.5 million repositories, documents, and enterprise sources. Owning the full domain-specific model lifecycle (data preparation, fine-tuning, deployment) and establishing AI governance frameworks driving organization-wide adoption.
Impact & Results
- ✓Process data across ~3.5M repositories, documents, and enterprise sources
- ✓Drive company-wide AI-first transformation as platform owner
- ✓Deliver critical AI systems from stalled state to production within aggressive timelines
- ✓Establish AI governance frameworks adopted across the organization
- ✓Lead New Relic's AI Community of Practice
Technologies Used
Challenge
XM platforms needed intelligent, conversational interfaces for analytics. Competitors like Qualtrics, QuestionPro, and Medallia were racing to build similar capabilities, creating a critical time-to-market challenge.
Solution
Built and shipped a production-ready gen AI powered chat assistant using LLMs, RAG architecture, and deep XM domain integration. Achieved market leadership through rapid iteration and customer-centric design.
Impact & Results
- ✓First to market before Qualtrics, QuestionPro, and Medallia
- ✓Won and onboarded Emirates Airlines as enterprise customer
- ✓Established competitive differentiation in the XM space
- ✓Reduced time-to-insight for analytics users by 70%
- ✓Enabled non-technical users to perform complex analysis
Technologies Used
Challenge
HappyFox needed to integrate AI across multiple products (Help Desk, Chat, BI) and unlock enterprise expansion — without fragmenting the architecture or accumulating technical debt.
Solution
Designed and built a centralized AI platform from the ground up. Delivered AI-powered systems including autonomous agents and intelligent assistants. Designed hybrid retrieval combining structured and unstructured data, optimized inference workflows for cost and efficiency, and supported enterprise sales motion (RFPs, customer solutioning).
Impact & Results
- ✓Drove 125% ARR growth through platform-led expansion
- ✓Built and scaled the AI platform from scratch, enabling multiple enterprise-grade products
- ✓Delivered autonomous agents and intelligent assistants in production
- ✓Designed hybrid retrieval systems combining structured and unstructured data
- ✓Supported enterprise sales, RFPs, and customer solutioning end-to-end
Technologies Used
Challenge
Xebo's legacy architecture couldn't support global enterprise deployments, the team needed to scale 4x, and the company had to clear multiple compliance bars simultaneously to enter regulated markets.
Solution
Rebuilt product architecture from scratch with multi-cloud + on-premise deployment options. Designed disaster-recovery setups across 4 regions. Served as acting CISO across SOC2, HIPAA, GDPR, PCI DSS, ISO. Drove automation and forward-deployment to compress onboarding from 6–12 months to 1 week–3 months.
Impact & Results
- ✓Drove 567% ARR growth in 24 months
- ✓Designed multi-cloud and on-premise architecture across 25+ enterprise deployments and 4 regions with DR
- ✓Achieved SOC2, HIPAA, GDPR, PCI DSS, and ISO compliance as acting CISO
- ✓Reduced enterprise onboarding from 6–12 months to 1 week–3 months
- ✓Featured in Gartner® Magic Quadrant for VoC Platforms
- ✓Led 35 engineers across R&D, Product, DevOps, and Compliance
Technologies Used
Challenge
AI models need robust evaluation frameworks to measure performance, accuracy, and safety across diverse scenarios. The community lacked standardized evaluation methodologies.
Solution
Developed evaluation scripts and test cases for the OpenAI Evals project, contributing to the open-source community's ability to assess LLM capabilities with standardized benchmarks.
Impact & Results
- ✓Recognized by OpenAI for valuable contributions
- ✓Helped improve evaluation methodologies for AI models
- ✓Contributed to open-source AI/ML ecosystem
- ✓Enabled better benchmarking across the industry