Seth DeWhitt 🎷

Software Engineer Jazz Musician Seattle, WA

If you have any questions about my experience, projects, or skills, start a chat!

HeadshotPlaying saxophonePicture of me!Presenting at a Jazz Education Network workshop
  • T

    The Data Mine, Purdue University

    Research Lead & Teaching Assistant | Practical AI Podcast

    • Directing 11 student researchers in collaboration with the Practical AI Podcast (1.5M+ listens) and Prediction Guard
    • Working under direction and mentorship of Dan Whitenack, CEO of Prediction Guard and a Practical AI host
    • Implementing a RAG chatbot with Perplexity-style citations trained on 349+ episode transcripts
    • Designed scalable podcast generation platform integrating LLM-produced transcripts and voice clones trained off podcast hosts
    • Providing technical guidance to team members, fostering a collaborative work environment
    • Planning/Leading biweekly Agile sprints, coordinating tasks and progress reviews across a 50k+ line codebase
  • C

    Chewy

    Software Engineer I Intern | Replenishment Engineering Team

    • Deployed scalable & containerized development environments with custom tooling for in-house model execution on AWS ECS/Fargate, reducing setup overhead by 80% and supporting 15+ engineers
    • Automated distributed system maintenance with AWS Lambda and Kinesis, lowering AWS expenses by 25%
    • Engineered persistent storage architecture linking Okta SSO authentication with AWS EFS, ECS, and RDS to support secure stateful development sessions across container restarts
    • Designed and implemented Terraform-based infrastructure provisioning and Jenkins CI/CD pipelines, decreasing deployment times by 66%
  • T

    The Online Marketplace Partnership (OMP)

    Intern | Data & Analytics

    • Integrated sales/advertising data from 5+ e-commerce platforms and 12+ databases into a unified KPI dashboard
    • Connected disparate sources including Amazon, Google, Meta, Walmart, Shopify, and WooCommerce
    • Enabled OMP experts to analyze data for 11 retail clients, with up to 60 performance metrics for 150 products
  • T

    The Data Mine, Purdue University

    Undergraduate Data Science Researcher | Bayer Crop Sciences, Syngenta

    • Trained a LightGBM model on 500k+ rows to predict climate-driven corn yield variation for Bayer clients at 91% accuracy
    • Optimized model performance through feature engineering and hyperparameter tuning, reducing RMSE by 24%
    • Implemented data preprocessing and QC pipelines for 100M+ data points, validated using mixed linear models

More About Me

I'm currently a Junior at Purdue University pursuing a B.S. in Computer Science with a concentration in Software Engineering. When I'm not coding, I'm likely practicing saxophone, arranging jazz music, or experimenting in the kitchen.

This past summer, I was a Software Engineering Intern at Chewy, where I pioneered a powerful, yet scalable service to provide data scientists with development environments loaded with custom tooling and essential libraries, reducing expenses spent on dry-running optimization/machine learning models on AWS during the development process.

Currently, I'm working on a browser extension to help users of all ages detect AI use in their social media feeds, helping them identify misinformation in an increasingly AI-driven world.

For more Q&A, start a chat or contact me!

Skills

Center logo
GitHub
Linux
Python
Node.js
Git
Docker
Bash
Next.js
React
TypeScript
PostgreSQL
TailwindCSS
Vite
Java
AWS
Jenkins
FastAPI
MongoDB
Supabase

Featured Projects

Here are some of my favorite projects :)

Evallm

Evallm

Web platform to assess performance of large language models

Next.jsVercelTypeScriptReactTailwindOAuth2MongoDBGroq API
RAG Portfolio Bot

RAG Portfolio Bot

A RAG chatbot built to answer questions about my experience, integrated into this website!

Next.jsVercelTypeScriptReactTailwindOpenAI APILangChainAI SDKSupabase

Blog

A peek into my life — inside and outside of software development

Current Favorite Songs

My sugar, spice, and everything nice