▰ FEATURED · PLAYER_01SESSION_2026 · 21:09:58:33

HARSHITHREDDYPETA

> engineer.exe ── ONLINE

I've worked across backend systems, AI tooling, and product-focused full-stack apps.

Recently I built an AI agent that compressed a 50-page welding manual into an interactive troubleshooting and setup assistant with multimodal support. I've also shipped a mobile app end-to-end that's live on the App Store, and built backend systems focused on reliability, performance, and real-world workflows.

One project that shaped how I think was at an Idea Fund hackathon, where I traced nearly $465K stuck in a billing flow back to a single broken state caused by fragmented systems across teams. Fixing it made me realize how important clean workflows and trustworthy systems are.

I like moving fast, owning product decisions, and building things people actually use.

IN QUEUE ◆↓ RESUME ↳ CONNECT

▰ READY

LVL 22

▰ PLAYER_01

Harshith

CS · UW–MADISON · 2026

INTERNSHIPS

3

shipped

GPA

3.8

dean's list 7×

CERTS

2

AWS · CP + DA

UPTIME

4 YRS

since 2022

◆TypeScript◆Next.js◆Python◆Spring Boot◆Postgres◆Firestore◆AWS◆Docker◆React Native◆GSAP◆fMRI◆LLMs◆TypeScript◆Next.js◆Python◆Spring Boot◆Postgres◆Firestore◆AWS◆Docker◆React Native◆GSAP◆fMRI◆LLMs

02 · CAREER

WHERE I'VE SHIPPED.

Roles are matches. Stack is a loadout. Achievements are badges. Pick a tab — they all hold up.

▰ RESEARCHMATCH #04

Undergraduate Research Assistant

UW–Madison · Computational Development Neuroscience Lab

Jun 2025 → Present · Madison, WI

Building reusable Python pipelines that turn raw fMRI and behavioral data into reviewable, subject-level reports.

▸Built a behavioral analysis framework processing 1,000+ experiment sessions across 5 cognitive tasks.
▸Automated 20+ behavioral and linguistic metrics, reducing recurring analysis time by 60%.
▸Developed an NLP pipeline using Word2Vec features with a Ridge / RBF-SVR / XGBoost ensemble, evaluated via 50-repeat nested split-half validation.
▸Implemented an fMRI reliability pipeline computing Cronbach's α and Guttman's G6 across Schaefer-atlas parcels.

Pythonscikit-learnXGBoostWord2VecfMRI

SESSIONS

1,000+

METRICS

20+

TIME SAVED

−60%

▰ STARTUPMATCH #03

Junior AI/ML Engineer Intern

SoundSafe.ai

May 2025 → Aug 2025 · Chicago, IL · Remote

Built shared audio preprocessing and verification services that standardized the ML pipeline across multiple detection models.

▸Developed a configurable SharedAudioPreprocessor (VAD, Mel-spectrogram, MFCC, RMS) that unified preprocessing across 4 downstream models and reduced cross-team data bugs by 50%.
▸Built a RESTful WatermarkVerificationService reaching 98%+ detection accuracy on test audio with processing latency under 750ms.
▸Partnered with the anomaly-detection team to integrate deepfake, anomaly, and behavioral models into the UTS pipeline, delivering sub-second inference and improving multi-event detection efficiency by 40%.
▸Co-designed the AudioFusionEngine, combining multi-model outputs via weighted fusion logic into a unified threat score; improved threat classification accuracy by 30% on benchmark datasets.

PythonPyTorchAudio MLREST APIsSignal Processing

DETECTION

98%+

LATENCY

<750ms

FUSION

+30%

▰ STARTUPMATCH #02

Full-Stack Software Engineer Intern

Take2

Mar 2025 → May 2025 · Remote

Improved app performance by redesigning Firestore access patterns and eliminating unnecessary React Native re-renders.

▸Redesigned Firestore queries and indexes, reducing retrieval latency by 40% and page loads by 1.5 seconds.
▸Resolved 10+ production issues across TypeScript components and Firebase security rules.
▸Restructured state flow in React Native screens, improving UI responsiveness by 35%.

TypeScriptReact NativeFirebaseFirestore

LATENCY

−40%

DEFECTS

−30%

FPS

+35%

▰ ENTERPRISEMATCH #01

Backend Java Developer Intern

Helixotech Pvt Ltd

May 2023 → Jul 2023 · Chennai, India

Built JWT-secured Spring Boot APIs for a healthcare platform managing 10K+ patient records.

▸Built JWT-secured authentication and authorization APIs using Spring Boot Security.
▸Refactored the service layer and added SQL indexes, improving patient-record retrieval efficiency by 20%.
▸Shipped 4 production REST APIs, validated by 15 automated Swagger integration tests.

JavaSpring BootJWTSQLSwagger

APIS

4

RECORDS

10K+

LOOKUP

+20%

03 · STORE