Artificial Intelligence × Engineering
Exploring the frontier where AI meets production systems. From LLM integration to Gen AI workflows — building intelligence into products that serve at scale.
Capabilities
Connecting large language models to real production systems. Prompt engineering, context management, and reliable output parsing for user-facing features.
Automated summarization, intelligent content recommendations, and AI-assisted editorial modules — shipped at scale on platforms with 69M+ users.
End-to-end generative AI pipelines for content automation. Building REST-based AI services and agentic workflows that integrate cleanly into existing stacks.
Deploying AI features in high-traffic production environments without performance regression. Optimized hydration, edge caching, and smart lazy-loading for AI modules.
Real-time data prioritization and intelligent monitoring dashboards for healthcare IoT platforms. Thousands of concurrent device records, rendered with zero lag.
Built and published the hindustan times.js SDK — enabling third-party platforms to embed real-time AI-powered news feeds. Designed for developer ergonomics and reliability.
Shipped Projects
Philosophy
AI should be
invisible to users
and invaluable
to the product.
The best AI integrations aren't flashy demos — they're seamless systems that make a product measurably better. Faster content discovery. Smarter recommendations. Zero added latency.
Building AI at the scale of Hindustan Times taught me that intelligence needs to be engineered with the same rigor as the underlying infrastructure. Performance is non-negotiable, even when you're running models at the edge.
Every feature I ship starts with a question: does this make the user's experience genuinely better? If the answer is yes, the engineering challenge becomes how to do it invisibly and reliably at millions of requests per day.