devesh.dev

CS @ UCSB

Devesh Mamidi

Software engineer building projects — from features powered by RAG pipelines to full-stack applications tested by real users. I am actively looking for a 2027 summer SWE internship

View My WorkDownload Resume
Headshot of Devesh Mamidi
Open to Opportunities

Find me

Projects

SentinelAI incident dashboard showing an active authentication incident with GitHub, Slack, and runbook status.

SentinelAI

~8s incident analysis
Try it outGitHub

Agentic incident response in ~8 seconds

AI-powered incident orchestration platform that performs semantic search over runbooks, correlates GitHub commits, and generates structured root-cause analyses with remediation steps.

  • RAG pipeline (FastAPI, ChromaDB, Gemini Embeddings) completes incident analysis in ~8 seconds.
  • Autonomous mode that uses Sentry to auto-detect bugs to trigger the pipeline
  • Full-stack React/TypeScript with Supabase auth, RLS, and Owner/Admin/Member RBAC.
  • Automated Slack postmortems aggregating AI analysis, GitHub evidence, and fix history into 9-section reports.
  • Containerized backend with Docker + persistent ChromaDB vector store across deployments.
  • Assign → fix → review → resolve workflow with role-aware React UI.
  • React
  • TypeScript
  • Tailwind
  • FastAPI
  • ChromaDB
  • Supabase
  • PostgreSQL
  • Docker
  • Gemini
  • Slack API
  • REST API

Where I've worked

  1. Jan 2026 – Present

    UCSB Caliber

    Web Developer

    Santa Barbara, CA

    Tested by 300+ students

    • Built a mastery-based course platform with assignment, practice, and instructor analytics workflows.
    • Delivered 6+ performance tracking views and summary statistics, tested by 300+ students.
    • Shipped 6+ features and 3+ bug fixes while resolving up to two issues per Agile Scrum sprint.
    • Collaborated with a 10+ member cross-functional team on a React-based course product.
    • React
    • TypeScript
    • Tailwind
    • Python
    • PostgreSQL
    • REST API

  2. Jan 2026 – Present

    Teach4Speech

    Web Developer

    Santa Barbara, CA

    66.7% → 95.8% accuracy

    • Built a three-level difficulty system for games by assigning 1,000+ words per difficulty in the backend.
    • Raised semantic word-matching accuracy from 66.7% to 95.8% with MiniLM sentence embeddings.
    • Wrote custom Node.js evaluation scripts to benchmark performance across labeled test sets.
    • Node.js
    • TypeScript
    • MiniLM
    • Tailwind
    • REST API

  3. Jul 2025 – Sep 2025 · Jun 2026 – Jul 2026

    TMAS Academy

    Web Developer

    Remote

    500+ students

    • Built a personalized AI chatbot for 500+ STEM AP students with quiz generation and study recommendations.
    • Integrated Gemini to dynamically create up to 10-question assessments for STEM AP topics.
    • Shipped a lightweight RAG system over 12 PDFs and 1,000+ practice problems by selected topic.
    • Next.js
    • Tailwind
    • pdf-parse
    • OpenRouter
    • REST API

What I
bring

Technical stack

Languages

  • Java
  • Python
  • C++
  • JavaScript
  • TypeScript
  • HTML/CSS
  • SQL

Frameworks

  • React
  • Next.js
  • Node.js
  • FastAPI
  • Flask
  • Tailwind

Databases

  • PostgreSQL
  • Supabase
  • Firebase
  • Convex
  • ChromaDB

Dev Tools

  • Git
  • GitHub
  • Docker
  • VS Code
  • Unix/Linux
  • Sentry

AI / ML

  • Gemini API
  • OpenRouter
  • MiniLM
  • RAG
  • Embeddings

Education

University of California, Santa Barbara

B.S. Computer Science

Class of 2029

3.80
GPA / 4.00
2x
Dean’s Honor List

Relevant coursework

  • Data Structures & Algorithms
  • Discrete Mathematics
  • Linear Algebra
  • Problem Solving with Computers

Leadership & involvement

  • UCSB Caliber

    Web Developer

    Agile Scrum, 10+ member team, tested by 300+ students.

  • Teach4Speech

    Web Developer

    Semantic NLP system, games difficulty architecture.

Let's work together

I'm actively looking for software engineering internships. Whether you have a role in mind or just want to connect — my inbox is open.