Devyn Duvall

Data Scientist & Full-Stack Engineer

I build production ML platforms, full-stack SaaS products, and cloud infrastructure that turn data into business impact — from 500M+-record pipelines to real-time IoT analytics.

View My Work Download Resume

About

Building systems that scale

I'm a data scientist and full-stack engineer based in Seattle. I design and build end-to-end ML platforms, SaaS products, and data infrastructure — from microservices architectures processing hundreds of millions of records to AI-powered applications serving real users in production. I care about clean code, thoughtful design, and shipping things that work.

500M+
Records processed
3
Production platforms built
13+
Microservices architected

Experience

Where I've worked

Lead Developer Jul 2025 – Present
Alpacka.ai
  • Launched a full-stack SaaS platform for outdoor gear management and AI-powered trip planning with 100 active users
  • Architected cloud-native infrastructure on AWS (ECS Fargate, RDS, ALB) with Terraform IaC across ~392 revisions
  • Built an ML-powered recommendation engine using OpenAI Assistants for intelligent packing suggestions
Founder & CTO May 2025 – Present
Capsul8.io
  • Designed and built a full-stack industrial IoT predictive maintenance platform with 60 React components and 54 backend services
  • Developed a 7-9 agent consensus anomaly detection system that reduces false positives by 60-80%
  • Built per-equipment ML models (Gradient Boosting, LSTM, Isolation Forest) with self-tuning configurations and continuous learning
Data Scientist Feb 2025 – Present
Deep Sync
  • Building a 13-service ML platform processing 500M+ records across a 13-step pipeline, training 360 XGBoost models with 1000+ engineered features and EMR-based batch prediction
  • Designed an enterprise QA framework validating 648+ fields across 15 data domains at billion-record scale on AWS EMR
  • Architected the full ecosystem: 10 repositories spanning microservices, event-driven Lambda orchestration, serverless reporting, and internal package distribution via CodeArtifact
Data Scientist Nov 2023 – Feb 2025
Fusion92
  • Increased marketing conversions by 265% and ROI by 500% through ML-driven audience segmentation
  • Managed a $400K data pipeline transformation for a Fortune 50 company using BigQuery, Looker, and Funnel.io
  • Drove $1.5M in annual revenue through automation of data workflows using Python, SQL, and REST APIs
Full Stack Developer May – Aug 2023
Westerly Kitchen
  • Overhauled the subscription creation process, driving a 13% increase in conversion rate
  • Orchestrated data migration with full-stack data pipelines, ensuring minimal disruption to business operations

Education

M.S. in Information Management, Data Science — University of Washington, Seattle (2023)

B.A. in Geography, Data Science Option — University of Washington, Seattle (2021)

Projects

What I've built

Capsul8.io

Founder & CTO — Industrial IoT Platform

A full-stack SaaS platform that transforms industrial IoT sensor data into predictive maintenance intelligence. The system connects to existing sensor infrastructure, applies three independent AI/ML subsystems, and predicts equipment failures hours to weeks in advance through multi-agent consensus voting.

7-9 Agent Consensus System 60 React Components 54 Backend Services Per-Equipment ML Models 3 AI/ML Subsystems
React Flask PostgreSQL TensorFlow LangChain AWS CDK Docker scikit-learn

Deep Sync ML Platform

Data Scientist — Enterprise ML Infrastructure

An enterprise-scale microservices ML platform orchestrating the full lifecycle of predictive analytics across 500M+ consumer records. The 13-service, 13-step pipeline trains 360 XGBoost models, runs EMR-based batch prediction with sophisticated post-processing, and includes an enterprise QA framework, change management service, and a full-stack React reporting dashboard with 50+ API endpoints.

500M+ Records 13+ Microservices 360 ML Models 1000+ Engineered Features 10 Repositories
Python XGBoost FastAPI Snowflake Snowpark PySpark AWS EMR Redis Docker

Alpacka.ai

Lead Developer — Outdoor Gear Platform

A full-stack SaaS platform for outdoor gear management and AI-powered trip planning. Users catalog equipment, organize storage, and receive intelligent packing recommendations powered by OpenAI Assistants. Deployed on AWS with Terraform IaC, ECS Fargate, and scheduled scaling for cost optimization.

~392 Terraform Revisions 100 Active Users ECS Fargate + RDS React Native Mobile App
React Flask PostgreSQL Terraform AWS ECS OpenAI React Native

Skills

Technologies I work with

Languages

Python JavaScript TypeScript SQL HTML/CSS

ML / AI

XGBoost scikit-learn TensorFlow LangChain OpenAI PyTorch Pandas NumPy

Cloud & Infra

AWS Terraform Docker AWS CDK ECS Fargate EMR Lambda CI/CD

Data Engineering

Snowflake PySpark Redis FastAPI Airflow ETL/ELT Databricks

Frontend

React React Native Tailwind CSS Vite Chart.js

Databases

PostgreSQL Snowflake SQLAlchemy BigQuery SQLite