Skip to content
View vidithsalla's full-sized avatar
1 follower · 0 following

Block or report vidithsalla

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
vidithsalla/README.md

Vidith Salla

Recent CS grad building end-to-end systems across backend, data, and AI. I am especially interested in problems where correctness, retrieval, and real-world constraints matter more than just getting something to "work."

Currently Building

  • Trace - a production-style document analysis system with grounded AI workflows, hybrid retrieval, and citation validation
  • Ventra - a multi-tenant construction ERP replacing Excel + WhatsApp workflows for procurement, inventory, and finance

Featured Projects

  • Trace - grounded document analysis, hybrid retrieval, citation validation, and production-style AI workflows
  • Ventra - construction ERP with multi-tenant workflows across procurement, inventory, and finance
  • Systems Programming Portfolio - pthread concurrency, memory allocator, scheduling simulators, and low-level systems work
  • Python DSL Parsers - lexers, recursive-descent parsers, ASTs, scoping, and type checking

Interested In

  • AI systems beyond demos: retrieval, grounding, and evaluation
  • Backend and data-intensive systems with real-world constraints
  • System design tradeoffs around correctness, latency, and reliability

Core Stack

Python TypeScript C SQL FastAPI Next.js React PostgreSQL Redis Docker GitHub Actions AWS GCP Supabase MQTT

Ask Me About

  • Building reliable RAG systems and what actually breaks in practice
  • Postgres, pgvector, async pipelines, and system design tradeoffs
  • Turning messy real-world workflows into software

Connect

Pinned Loading

  1. trace trace Public

    Full-stack document analysis system with hybrid retrieval (FTS + vector + RRF) and deterministic citation validation to ensure grounded AI outputs.

    Python

  2. checkgraph checkgraph Public

    Event-driven restaurant check operations and funding-reconciliation prototype for payments, hosted deposits, coverage allocation, and support recovery.

    TypeScript

  3. harper-marketfit-workbench harper-marketfit-workbench Public

    Unofficial commercial insurance submission workbench prototype with evidence backed extraction, market routing, review traces, and evals.

    TypeScript

  4. chartapilot chartapilot Public

    Synthetic urgent care AI chart review deployment workbench with evidence-backed findings, human review queues, audit logs, and evals.

    Python

  5. voice-agent-deployment-lab voice-agent-deployment-lab Public

    TypeScript

  6. c-operating-systems-projects c-operating-systems-projects Public

    C