benschwemlein.com

Ben Schwemlein

Senior Software Engineer

RESUME Senior Software Engineer

Senior software engineer with 15 years building Java and full-stack systems. Eight of those as a consultant, delivering for clients across healthcare, retail, food service, and logistics, including leading a development team and owning the architecture on a greenfield platform. An early adopter of AI-assisted development. AWS Certified Developer with an M.S. in Computer Science from DePaul University.

Skills

  • Java
  • Spring Boot
  • Angular
  • TypeScript
  • Python
  • AWS
  • Google Cloud
  • Microservices
  • Kafka
  • Docker
  • SQL and NoSQL
  • AI assisted development
  • Tech lead

Experience

Leading EDJEFeb 2018 – Aug 2026 Senior Software Engineer. Consultant delivering full-stack and cloud solutions for clients. Clients included Jumpmind, Safelite, Donatos Pizza, Crown Equipment Corporation, Cardinal Health, DSW / Designer Brands, Wendy's International.
Nationwide Insurance2015 – 2018Senior Java Developer on enterprise applications for the retirement-plan business.
OCLC2013 – 2015Software Engineer on OCLC.org and its content-management system.
Nationwide Insurance2011 – 2013Senior Java Developer on the Affinity program web service and a call-center application for insurance policy inquiries.

Education

DePaul UniversityChicago, IL M.S. Computer Science, with honors (independent study: machine learning & NLP research). Upsilon Pi Epsilon.
Ohio Wesleyan UniversityDelaware, OH B.A. Economics.

Certification

AWS Certified Developer – Associate (2024)

Download the full resume (PDF)

BLOG Notes and write ups

1 / 3

PROJECTS Side work

LocalScope
Local AI code intelligence. Semantic search and question answering over any codebase with Ollama and ChromaDB, nothing leaves your machine.
radio_telescope
Hobby radio telescope: an RTL-SDR dongle, a Raspberry Pi, and a PyQt6 planning app, aiming at a hydrogen line map of the Milky Way.
oblivion-landscape-viewer
Renders Oblivion landscape meshes in 3D. The same viewer written in Rust, Python, Kotlin, and two flavors of Java, as a language comparison.
library-catalog-app
A Spring Boot and Angular library catalog, kept as a sample app and a test bed for LLM coding tools.
Naive-Bayes-Text-Classifier
Naive Bayes document classifier from my machine learning course at DePaul.

All of it on GitHub.

CONTACT Say hello

Open to senior engineering roles, remote or around Columbus. Email is the fastest way to reach me.

ABOUT ME Ben Schwemlein

Ben, rendered as a cyan hologram on a curved CRT screen
Name
Title
Location
Industries
Skills
Certified
Education
Hobbies

My thoughts on the anniversary of Challenger

Thirty years after the Challenger disaster, what a fourth grader remembers about that morning, and what Christa McAuliffe still stands for.

The Teacher in Space mission patch, an embroidered NASA patch with the shuttle rising over Earth
The Teacher in Space mission patch.

Thirty years ago this week, on January 28, the Challenger shuttle disintegrated in a disaster that shocked most people in a way that those who were not around to experience it may not fully appreciate. The loss of the shuttle and the first teacher in space, Christa McAuliffe, was more painful than other disasters such as airline crashes. The space shuttle represented the pinnacle of American technology in a time before the internet, when high tech meant big engineering, and when our nation was competing against the Soviet Union for dominance. Christa McAuliffe, chosen out of ten thousand teachers, had an infectious enthusiasm for learning and a hope for the future that seemed to represent the positive essence of 1980s United States. The space shuttle had been marketed as an airliner to space that would open the final frontier to common Americans, and Christa was the first civilian, the first one of us, who was going for the ride. Her personality was magnetic and she captivated us with her journey. My fourth grade teacher had a poster of her in our classroom.

I first learned of the disaster when my teacher walked into the room in tears. She said she had some sad news, and told us, with the deliberate sensitivity with which one announces death to children, that the space shuttle had exploded and the Teacher in Space had died. As a nine year old I tried to comprehend what I had been told. But I knew my teacher was hurt because another teacher, one of her own, had died. Later we watched the explosion on TV. The sky was a clear brilliant blue up to which the shuttle's cotton white exhaust climbed like a Greek column. The white column climbed and climbed like Jack's beanstalk and then crumbled in an eerie silent tangerine colored puff. That puff was hard to watch because we witnessed the instant loss of Christa and the crew. In a broader sense it represented the crumbling of some of the hope that was part of the 80s.

Even after all these years Christa's energy reminds me how central education, teachers, and technology are to building a hopeful future. I think of her and am reminded of how wonderful my children's teachers are. I think of her journey into space and am inspired to use technology for a better future. I think of how my teachers nurtured my happy childhood in the 80s. I am reminded how much I respect and am humbled by my friends who are teachers or professors.

If you haven't yet, watch Christa talk on YouTube. See if you're inspired as others were.

Originally posted on LinkedIn.

99 Problems in Scala

99 Problems teaches Scala by solving common problems, mostly with a functional approach. Problem 8 shows off pattern matching.

99 Problems teaches Scala through solving common problems, mostly using a functional approach. Give them a try.

The solution to problem 8 illustrates Scala's pattern matching by using its case syntax.

// P08 (**) Eliminate consecutive duplicates of list elements.
//     If a list contains repeated elements they should be replaced with a
//     single copy of the element. The order of the elements should not be
//     changed.
//
//     Example:
//     scala> compress(List('a, 'a, 'a, 'a, 'b, 'c, 'c, 'a, 'a, 'd, 'e, 'e, 'e, 'e))
//     res0: List[Symbol] = List('a, 'b, 'c, 'a, 'd, 'e)

object P08 {
  // Standard recursive.
  def compressRecursive[A](ls: List[A]): List[A] = ls match {
    case Nil       => Nil
    case h :: tail => h :: compressRecursive(tail.dropWhile(_ == h))
  }
}

Originally posted on benschwem.wordpress.com.

SS Cygni, my favorite (binary) star

A variable binary star 400 light years away that erupts in a dwarf nova about every 50 days, and the amateurs who keep watch on it.

Before the birth of my four children, I spent many relaxing evenings observing an interesting star system through my telescope. SS Cygni is a variable binary star system approximately 400 light years from Earth. It is composed of two stars swinging around each other in a close mutual orbit, with one gravitationally pulling hot gas off the other's surface. This hot gas forms an accretion disc of plasma around the heavier star, which erupts in a tremendous explosion, called a dwarf nova, about every 50 days. The heavier star, a white dwarf, has a much smaller diameter and is much more dense than its companion red dwarf.

Artist's view of a cataclysmic binary star with an accretion disc
A cataclysmic binary similar to SS Cygni. The denser primary and its accretion disc are on the right, the secondary on the left.

At 400 light years, SS Cygni appears as a barely detectable dim dot in a telescope, which pushes your eyes to their limits. With a little patience and practice you can witness the eruption as a 40 fold increase in the brightness of that dot. Its brightness is measured by comparing it to nearby stars. Since the eruptions are unpredictable, professional astronomers rely on amateurs, such as myself, to post these measurements to the American Association of Variable Star Observers. The aggregate measurements from around the world form a plot of the star's brightness over time.

Light curve of SS Cygni from 2005 with sharp spikes at each eruption
The magnitude over time of SS Cygni in 2005. The spikes are eruptions of the accretion disc.

Taking measurements is tedious and lugging a 12 inch telescope is back breaking, but there was a spirituality in having my daily routine synchronized with events over 400 light years away, with energies dwarfing anything we experience on Earth. Like our awe at Old Faithful, SS Cygni is a pleasure of creation that only we humans truly appreciate experiencing. Of all life's pleasures, the curiosity that draws us to places like SS Cygni reminds me that the unique experience of being human is wonderful.

Originally posted on benschwem.wordpress.com.

CLOUD AND AI In detail

Cloud

AWS Certified Developer, Associate (2024). Production work on AWS, Google Cloud, and Pivotal Cloud Foundry.

  • Wendy's Built a high volume AWS platform in two months serving point of sale data from 6,000 restaurants to 13 vendors, replacing a legacy file transfer system. Chose Aurora MySQL and wrote the CloudFormation for it, then re-architected onto S3 mid build when load testing exposed a scaling limit. Containerized the service for Docker and Kubernetes, set up logging through the ELK stack, and designed the certificate based authentication for devices and vendors, including revoking certificates from individual devices, taking it through the client's security reviews.
  • Safelite Recommended AWS CodeDeploy after showing the initial SSM and EventBridge approach would break under EC2 Auto Scaling. Stood up the CodeDeploy applications, deployment groups, and lifecycle hooks wired to Azure DevOps, ported the release steps off the licensed tool, cutting roughly $70K a year, documented blue green and canary rollout options with load balancers and Auto Scaling, and updated the Windows Server images for predictable provisioning.
  • Donatos Pizza Alongside an AWS architect, hardened a revenue critical environment: fixed high severity findings from a security audit, resolved drift between CloudFormation and manual changes, restructured the templates into modules, and used import to stage and validate changes before production. Evaluated Terraform as an alternative.
  • Cardinal Health Delivered on Google Cloud and Pivotal Cloud Foundry. Designed the Google Cloud Storage integration for invoice documents from Cloud Foundry, presenting cost estimates and security options to the architects, a solution three other teams adopted. Worked with infrastructure engineers on hosting MongoDB on Compute Engine.
  • Leading EDJE A serverless AWS service on API Gateway, Lambda, Bedrock, and Aurora, and an AI incident system running on EKS with DynamoDB and CloudFront.
  • Crown Equipment, DSW Java microservices on AWS behind the fleet management dashboard, and Google Cloud with Firestore for the rewards platform.

AI

Building with AI since 2024, and machine learning and NLP research before that.

  • Leading EDJE, 2024 Designed and built a serverless AWS service using LLMs on Bedrock to summarize and manage knowledge from the firm's technical Slack channels, on API Gateway, Lambda, and Aurora in Python.
  • Donatos, 2025 Used AI to generate the command line work that reproduced the infrastructure drift in an isolated account so template changes could be verified safely.
  • Jumpmind, 2025 to 2026 Built offline semantic code search over a large proprietary point of sale codebase with Python, ChromaDB, and Ollama: embeddings for retrieval and a local model for question answering, because the client's source could not leave their network. Ahead of similar tooling the team adopted later, and measured against a labeled question set so retrieval changes were chosen on evidence.
  • Leading EDJE, 2026 Part of the team behind an AI driven incident detection and remediation system on AWS, with agents that diagnose production failures and open pull requests with fixes. Built the demonstration workload, a diagnosis skill for its flow engine errors, and the dashboard.
  • Leading EDJE, 2026 Built a Claude Code skill that turns AI generated feature plans into Jira epics and tasks with a preview and confirm step. Ran three technical spikes in half a day and generated a feature's thirteen implementation tickets in a few hours with the same workflow.
  • Open source LocalScope on GitHub: local code intelligence with semantic search and question answering over any codebase, with an evaluation suite.
  • DePaul University Funded research using machine learning and NLP to extract conclusions from biomedical journal articles, and a Naive Bayes document classifier from the machine learning course.

SKILLS Technical skills

Core

  • Java
  • Spring Boot
  • Angular
  • TypeScript
  • JavaScript
  • HTML5
  • CSS
  • Python

Architecture

  • Microservices
  • REST APIs
  • Event-driven / messaging (Kafka)
  • Enterprise integration (SAP, ServiceNow, Dell Boomi)

AWS

  • Lambda
  • API Gateway
  • S3
  • Aurora
  • ECS
  • EC2 Auto Scaling
  • CloudFormation
  • CodeDeploy
  • CloudWatch
  • Bedrock
  • DynamoDB
  • IAM

Cloud / DevOps

  • Google Cloud Platform
  • Pivotal Cloud Foundry
  • Docker
  • Kubernetes (EKS deployments)
  • CI/CD (GitLab CI, Azure DevOps)
  • Gradle
  • Maven
  • Git

Data

  • JPA
  • MongoDB
  • MySQL
  • Oracle
  • PostgreSQL

Practices

  • Agile / Scrum
  • Kanban
  • code reviews
  • mentoring

AI-Assisted Development

  • Claude Code
  • agent skills
  • SpecKit workflows
  • offline RAG and semantic search (embeddings, ChromaDB, Ollama)
  • retrieval evaluation
  • LLM-as-judge

EXPERIENCE 2011 to 2026

Leading EDJEFeb 2018 – Aug 2026 Senior Software Engineer. Consultant delivering full-stack and cloud solutions for clients.

Internal AI initiatives

  • Part of a team that built an AI-driven incident detection and remediation system on AWS, with agents that diagnose production failures and open pull requests with fixes
  • Built a custom skill that turns AI-generated feature plans into Jira epics and tasks, extending a SpecKit AI development workflow that produces feature specs, technical plans, and data models
  • Built a serverless AWS service using LLMs to summarize the firm's technical Slack channels (API Gateway, Lambda, Bedrock, Aurora, Python)

Claude Code, SpecKit, Amazon Bedrock, DynamoDB, Lambda, API Gateway, CloudFront, EKS, Java, Spring Boot, Python

Jumpmind Retail point of sale platform provider

  • Delivered a new store credit card application flow for American Eagle that let customers apply outside of a purchase, defining the reporting and application tracking requirements with stakeholders
  • Built offline AI semantic code search (Python) for a large proprietary codebase, no source leaving the client network, measured against a labeled question set so retrieval changes were chosen on evidence, not instinct

Java, Spring Boot, Angular, TypeScript, SCSS, PostgreSQL, Python, ChromaDB, Ollama

Safelite AWS deployment pipeline modernization

  • Modernized the deployment pipeline into AWS-native tooling, cutting roughly $70K in annual licensing costs
  • Recommended and built the pipeline using AWS CodeDeploy over a more brittle approach

AWS CodeDeploy, EC2 Auto Scaling, CloudFormation, CloudWatch, Azure DevOps, CodeBuild

Donatos Pizza Cloud security and infrastructure hardening

  • Hardened a revenue-critical AWS environment by fixing high severity security findings and resolving CloudFormation drift, reproducing it in an isolated account to verify template changes safely

AWS (IAM, CloudFormation, EC2, RDS, Route 53, Certificate Manager, ELB), Wiz, Python, Git

Crown Equipment Corporation Fleet-management and dealer-system modernization

  • Built new microservices within a large-scale architecture to power a customizable fleet-management dashboard
  • Modernized legacy dealer/partner systems onto a current Java/Angular cloud stack

Java, Spring Boot, Angular, RxJS, TypeScript, JPA, Flyway, RabbitMQ, AWS, Docker, MySQL, Oracle, Gradle, Maven, Jasmine

Cardinal Health Greenfield logistics invoicing platform

  • Tech lead over a team of 6, owning the invoicing platform's architecture and integrations and mentoring its developers
  • Designed the platform's Kafka integrations, intake and publishing, and worked out the data contracts directly with the upstream and downstream teams
  • Designed a document-management service and a realistic-data load-testing seeder, both adopted by other teams in the program

Java, Spring, Angular, MongoDB, Kafka, Google Cloud Platform, Pivotal Cloud Foundry, Okta, Dell Boomi, ServiceNow, Apigee, RxJS, Mockito

DSW / Designer Brands Customer rewards and promotions platform

  • Sole developer for the coupon campaign management system, taking it from requirements with a business analyst through production release

Java, Spring, Angular, RxJS, Google Cloud Platform, Firestore, MySQL

Wendy's International High-volume point-of-sale data platform

  • Built a high-volume AWS platform serving point-of-sale data from 6,000 restaurants to 13 third-party vendors, replacing a legacy file-transfer system
  • Re-architected the project mid-build when load testing exposed a scaling limit, keeping the project on track
  • Implemented the platform's certificate-based authentication for devices and vendors, including revoking certificates from individual devices, and took it through the client's security reviews

Java, Spring Boot, JUnit, Docker, Gradle, GitLab, MySQL, AWS Aurora, CloudFormation, S3, S3 Java SDK, ELK

Nationwide Insurance2015 – 2018 Senior Java Developer on enterprise applications for the retirement-plan business. Java, Spring, Angular, JAX-RS, Oracle, Websphere
OCLC2013 – 2015 Software Engineer on OCLC.org and its content-management system. Java, Adobe CQ, JavaScript, jQuery, Maven, JUnit
Nationwide Insurance2011 – 2013 Senior Java Developer on the Affinity program web service and a call-center application for insurance policy inquiries. Java, Spring MVC, Grails, Groovy, jQuery, Oracle, DB2, Websphere, SOAP

EDUCATION And certification

Certification

AWS Certified Developer – Associate (2024)

Education

DePaul UniversityChicago, IL M.S. Computer Science, with honors (independent study: machine learning & NLP research). Upsilon Pi Epsilon.
Ohio Wesleyan UniversityDelaware, OH B.A. Economics.

HIGHLIGHTS Four things

Tech lead and architecture
Led a team of four to six on a greenfield invoicing platform for almost three years, owning its architecture and the integration design across the teams it connected to.
Cloud
AWS Certified Developer. Built, deployed, and hardened production systems on AWS, and delivered on Google Cloud and Pivotal Cloud Foundry.
AI
Machine learning and NLP research at DePaul, then building with LLMs since 2024: a Slack knowledge service on Bedrock, offline semantic code search, incident remediation agents, and Claude Code as a daily development workflow.
Full stack delivery
Java and Spring on the back end with Angular in front, from point of sale systems to fleet management to insurance.
NAV
HOME
1 / 6
Drag to turn, arrow keys work too. F fills the screen with a face.