Network Growth AI Infra 🚀

• Tech Lead building AI Infrastructure at LinkedIn scale: from LLM productionalization to real-time GPU based inference serving 1B+ members

• Led end-to-end ML infra for Embedding-Based Retrieval (EBR), Candidate Generation, and Deep Learning ranking systems powering People You May Know

• Shipped fine-tuned LLM embedding models into production: offline batch pipelines, online real-time serving, and KNN-based retrieval at scale

• Passionate about the intersection of large-scale distributed systems and applied AI: turning research models into production infrastructure that moves metrics

• Built, led and owned multiple 0 to 1 products such as Contextualization for AI Candidate Generation, Next Best Edge Model Scoring Infra, etc.

Feed Discovery Charter Tech Lead

• Leading a team of engineers across stacks to re-architect the Infrequent Member Feed Experience

• Eng designed and implemented onboarding and feed experiences that focus on edge building on LinkedIn Feed AI and Feed Infra ecosystems including Content by Follow Recommendations

• Created a new Feed Relevance paradigm to rerank content based on personalized user templates and behavior; leveraged this in the first Feed AI models with ML based UX on feed updates

• Co-Tech lead for the introduction of the Team Moments content type to Feed, focused on API and backend integration

Feed Frontend and AI Eng

• Built out the Hashtag ecosystem at LinkedIn to create member engagement in topical communities. Migrated old topics to hashtags, built member onboarding experiences

• Restructured novice member feed by removing promotional content from their Feed AI models. Built out this experience on the infra, ML and UI stack

• Led the Follows migration effort on frontend across pillars

• Improved Feed experience for members by adding customization of Newsfeed content

• Collaborated with iOS/Android Clients and Backend services to create a relevance sorted Follow Hub with curated packages where members can follow Recommended Members, Companies and Channels, and an Unfollow Hub to similarly unfollow suggested non-relevant Connections

• Created REST endpoints to allow developers to test and debug Feed updates

• Worked on the Autodesk Developer Network team and created a Ruby SDK for the View and Data API, to allow easier embedding of the Autodesk LMV

• Published a Ruby Gem to allow usage of the API in Ruby scripts and RoR apps, as well as a simple tutorial app and a sample competitive prototyping app

• Created JavaScript extensions for the A360 Viewer to facilitate better viewing of content and added tools for specific tasks such as edging closer to or rotating a model

• As an Autodesk Developer Evangelist, conducted the Angel Hacks Silicon Valley Workshop. Described the View and Data API and gave a tutorial on setup of the LMV to over 50 developers in attendance. Tabled at the hackathon, explained usage of the API and helped out participants with their implementations