Hoard
LiveContextual bookmark manager & creator studio for articles, videos, repos, and papers — with semantic search, multi-part series roadmaps, Kanban pipeline, PWA offline caching, and a Manifest V3 browser extension.
Software Engineer at Rivian and full-stack AI engineer with 4+ years of production experience. I build end-to-end LLM applications, semantic search & RAG pipelines, and scalable web platforms.
Projects built and shipped in 2026. Click any pin on the release timeline to jump directly to that project.
Contextual bookmark manager & creator studio for articles, videos, repos, and papers — with semantic search, multi-part series roadmaps, Kanban pipeline, PWA offline caching, and a Manifest V3 browser extension.
Daily learning campaigns: AI-generated study plans up to 730 days, spaced repetition on a 7/30/90/180 ladder, XP and badges, and a Field Kit for notes and bookmarks.
Interactive debt payoff accelerator, strategy simulator, payment tracker, and financial analytics dashboard with offline PWA support.
Distributed Redis-backed rate limiter supporting 5 algorithms (Token Bucket, Sliding Window, Fixed Window, Leaky Bucket, Composite) with adaptive traffic shaping and 472+ tests.
Turn any book summary into 3 actionable habits and a personalized 30-day plan, powered by Google Gemini.
Generative-AI concepts broken down and explained — interactive visualizers for embeddings, attention, RAG, and agent loops.
Structured resume parsing, JD alignment scoring, and bullet-point rewrite suggestions tailored to target engineering roles.
Resume feedback with the politeness filter removed. Built and shipped over a single weekend.
This site — built with Next.js 16, React 19, live GitHub activity integration, Gemini recruiter tools, and a custom neobrutalist design system.
Self-learning platform that generates structured multi-phase course plans, hands-on projects, and persisted checkpoints from any learning goal using Google Gemini.
Early prototype of bookmark capture with an AI tagging layer on top — evolved into Hoard.
A job board and application tracker built for developers rather than recruiters. Seven weeks of commits from first push to last.
Inspect live commits and core modules across my latest 2026 builds.
Add multi-part series paste import and parsing capabilities
Redesign studio editor top bar and layout structure
Enhance studio search and filter with roadmaps collapsible drawer
Add StudioOmnibar component and modal views for studio interface
Set default series view state to collapsed for all series
Vector embeddings + hybrid search across articles, papers, videos, and repos with d3-force graph view
Kanban drag-and-drop board, multi-part series importer, cover generator, and AI script drafting
One-keystroke tab saving (Alt+Shift+H), context-menu capture, and instant popup search
Offline Service Worker caching and native share sheet integration on iOS, Android, and desktop
Enterprise software engineering across EV & energy platforms, distributed cloud backends, and production AI/LLM systems.
Building scalable software systems, backend services, and full-stack web applications across Rivian's EV and energy platforms.
Architected production RAG search pipelines on AWS Bedrock (+15% ecommerce conversion across a 10M+ catalog), built LangGraph multi-agent workflows (-60% manual finance triage), and optimized Next.js storefronts (-40% bundle size).
Designed and maintained distributed backend services, REST APIs, relational data models, and cloud infrastructure across production deployments.
755 contributions in the past year · @srimaniteja19
Technologies used across enterprise systems at Rivian and independent full-stack AI products.
Server-first React architectures, streaming UI, and type-safe client state.
Production RAG pipelines, hybrid vector/keyword search, reranking, and multi-step agents.
Relational schema design, vector indexing, low-latency caching, and transactional integrity.
Containerized microservices, CI/CD automation, payment webhooks, and observability.
From initial problem framing to shipping and iterating in production.
Clarify user and system constraints, define latency and accuracy targets, and scope the smallest high-value release.
Model relational and vector schemas, API contracts, and retrieval pipelines before writing application glue code.
Ship type-safe TypeScript, React Server Components, and tested backend services incrementally through CI/CD.
Measure production p95 latency, retrieval quality, and real user behavior to guide the next iteration.
Always open to interesting engineering problems and collaborations. Feel free to reach out for roles, open-source collaboration, or engineering conversations.
Open to connecting on: