Open to Summer 2027 internships

Hi, my name is

Achintya Rai.

I build backend systems that move fast.

I'm a Computer Engineering major at UC San Diego focused on backend systems, distributed systems, performance, and low-latency computing. Off the clock, I practice Taekwondo and play competitive chess.

View my work Get in touch
Achintya Rai
Scroll
0 Payment events profiled
0 Research pipeline speedup
0 IMDb entries processed
0 GPA at UC San Diego

01. About

I'm a second-year Computer Engineering student at UC San Diego (GPA 3.9), drawn to the parts of software that have to be fast, correct, and hold up under load — data pipelines, distributed systems, and the machine-learning infrastructure around them.

Recently I've been an SWE intern at Adobe building time-series anomaly detection over 60M+ payment events, a founding engineer at VALT Health, and an NSF REU research fellow studying the geometry of LLM embeddings.

A few things I've been working with lately:

  • Python
  • C / C++
  • TypeScript
  • PyTorch
  • Django / FastAPI
  • AWS / GCP
Studying
B.S. Computer Engineering, UCSD
Graduating
June 2028
Focus
Backend · Distributed Systems · ML Infra
Location
San Diego, CA

02. Experience

Software Engineering Intern · Adobe

June 2026 – Sept. 2026 · San Jose, CA

  • Profiled 60.6M+ payment events in Databricks, surfacing 86,387 duplicate records and a 47% null rate in timestamp fields to establish the correct source of truth for downstream monitoring.
  • Identified a 26-hour processing outage via hourly volume analysis across 1,272 payment-monitoring segments, making it the primary validation benchmark for the anomaly detector.
  • Designed a time-series anomaly detector, iterating over statistical and ML models to handle multi-layer seasonality across 10+ workflows and 30+ countries.
  • Built SQL and MLOps pipelines for weekly retraining and hourly inference across all forecasting models.
  • Databricks
  • Python
  • SQL
  • MLOps
  • Time-Series

Founding Software Engineer · VALT Health

Aug. 2025 – Feb. 2026 · Remote

  • Engineered a Flask backend handling 100k+ datapoints from fitness devices (Fitbit, Apple Watch, etc.), designing schemas to unify data formats across 20+ distinct sources.
  • Implemented async batch processing and storage workflows in GCP, yielding a 35% speedup and 60% fewer redundant database writes.
  • Automated local environment setup and data seeding via Bash scripts, cutting testing time by up to 80%.
  • Authored technical specs and documentation standards to reduce onboarding friction for new engineers.
  • Flask
  • GCP
  • Python
  • Bash
  • PostgreSQL

Software Engineering Intern · Decklar (formerly Roambee)

June 2024 – Aug. 2024 · Sunnyvale, CA

  • Developed real-time shipment-monitoring modules for Amazon, Coca-Cola, and Samsung, optimizing Python/NumPy pipelines for a 15% capacity boost on 1M+ weekly events.
  • Implemented advanced data structures for visualization and analytics modules, yielding 20% faster reporting and better data integrity across internal dashboards.
  • Improved workflow efficiency 15% through Panels-based UI/UX refinements and iterative user testing.
  • Python
  • NumPy
  • Panels
  • Data Viz

03. Research

Undergraduate Research Fellow, NSF REU · Graph Theory & Neural Architecture

March 2026 – Present · San Diego, CA

  • Engineered a Python pipeline to analyze LLM embedding manifolds (Llama-2, DeepSeek-V2), reducing data footprints from 500GB checkpoints to 13GB localized manifolds.
  • Optimized placement scripts and preprocessing for a 30× runtime reduction (30 min → <1 min), enabling large-scale manifold analysis on 8GB-RAM systems.
  • Applied Rent's Rule and Louvain clustering to identify model "backbones," finding noise-filtering raises Inter-Cluster Network density from 0.18% to 13.25% in Llama-2-70B.
  • Python
  • LLMs
  • Graph Theory
  • Manifold Analysis

Undergraduate Researcher · Data Science Teaching & Learning Lab

Sept. 2024 – Present · San Diego, CA

  • ChatSight — built an LLM-augmented data-wrangling tool (FastAPI + Gemini API) that auto-labels 16k+ datapoints of student AI usage, cutting labeling time 90% over three UX iterations.
  • WISE Jupyter AI Tutor — architected a multi-agent tutoring framework with hallucination-reduction strategies for 95%+ response accuracy, synthesizing 100+ notebook cells and 100k+ datapoints in real time.
  • FastAPI
  • Gemini API
  • OpenAI API
  • Multi-Agent

Undergraduate Research Fellow, UC Scholars · Multi-Agent Systems

June 2025 – Aug. 2025 · San Diego, CA

  • Built a multi-agent AI chess tutor (Django + React + SQLite) via Agile sprints, cutting hallucinations 80% with an 83% faster runtime.
  • Developed 7 REST endpoints and production prototypes validated by 15+ chess players and coaches.
  • Published and presented at the UCSD Summer Research Conference (SRP).
  • Django
  • React
  • Gemini API
  • SQLite

04. Projects

Watchdog

April 2026

An agentic browser safeguard using the browser-use SDK and Gemini Vision to autonomously monitor tabs and intervene on unproductive sites via GUI automation, with a concurrent Asyncio CLI and Notion task sync.

  • Python
  • Gemini API
  • browser-use
  • Asyncio
  • SQLite

NanoGPT

March 2026

A GPT-style transformer built from scratch in Python/PyTorch — multi-headed attention, causal masking, and backprop layers — with critical sub-modules compiled in C++ for 25% lower per-token latency.

  • PyTorch
  • C++
  • Transformers
  • LLM Optimization

SquaredUpChess

Aug – Dec 2025

A full-stack platform on AWS for coach–student scheduling, real-time lesson tracking, and payments. React frontend, 20+ Django REST endpoints, and a CI/CD pipeline for 99.9% deployment reliability.

  • React
  • TypeScript
  • Django
  • WebSockets
  • AWS
  • Docker

MovieRecapinator

April 2025

A Spring Boot recommendation engine processing 700,000+ IMDb entries to surface high-priority episodes for series recaps — built with a 4-person team at DataHacks in a 24-hour sprint.

  • Java
  • Spring Boot
  • Maven
  • IMDb API

05. Skills

Languages

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

Development

  • React
  • Django
  • Node.js
  • FastAPI
  • Asyncio
  • WebSockets
  • Flask
  • JUnit
  • PyTest

AI & Cloud

  • PyTorch
  • AWS
  • GCP
  • Gemini / OpenAI API
  • Docker
  • GitHub Actions
  • Linux

Checkmate

06. Get in touch

I'm always happy to talk about backend systems, research, chess, or internship opportunities. My inbox is open — make the first move.

Say hello