# Colton Hester > UC Berkeley MIDS Student | ML Engineer | Specializing in AI Safety & Trustworthy Systems Location: San Francisco Bay Area, United States Profile: https://flows.cv/coltonhester Graduate student in UC Berkeley's Master of Information and Data Science (MIDS) program, building toward a career in AI safety and alignment research. Currently applying ML and data engineering to healthcare systems at Contra Costa County Health Services while developing expertise in causal inference, interpretability, and trustworthy AI systems. My technical foundation spans 7+ years of full-stack development and cloud architecture (4x AWS certified), which I'm now leveraging to address critical challenges in AI reliability and safety. Current coursework includes Machine Learning at Scale and Neurosymbolic AI, building on prior work in Applied Machine Learning and Experiments and Causal Inference. Recent research: Designed experimental study investigating emergent misalignment in LLM fine-tuning, extending work by Betley et al. on how narrow training can produce broad behavioral changes - a core question in AI safety. Core interests: AI alignment, interpretability, neurosymbolic architectures, and building ML systems that are reliable, transparent, and beneficial. Particularly drawn to the challenge of creating AI systems that remain trustworthy as they scale in capability. Actively seeking: Summer 2026 internships in ML engineering, AI safety research, or applied ML roles where I can contribute to building more reliable and interpretable AI systems. ## Work Experience ### HSIS Programmer & Analyst @ CONTRA COSTA COUNTY HEALTH SERVICES Jan 2024 – Present | Concord, California, United States Core Responsibilities: - Led end-to-end development of Behavioral Health Bed Management system: served as lead developer responsible for requirements gathering, system architecture design, and stakeholder coordination; created comprehensive scope documentation and presentation materials for IT Steering Committee (ITSC) approval process; managed cross-functional collaboration between clinical staff and IT teams to define workflows for bed capacity tracking and patient placement optimization - Design and implement healthcare compliance solutions, including Transparency in Coverage data pipeline automation using Azure Blob Storage and automated JSON file synchronization - Maintain and modernize legacy healthcare systems, troubleshooting complex .NET applications, SSIS packages, and SQL Server integrations - Develop PowerShell and CLI automation scripts for ETL workflows, reducing manual processing time and improving data accuracy - Collaborate with cross-functional teams to gather requirements and deliver technical solutions for healthcare data management initiatives - Ensure HIPAA compliance and data security protocols across all development projects Key Achievement: - Resolved 2+ year backlog in regulatory compliance reporting through automated solution implementation, ensuring organizational adherence to federal healthcare transparency requirements Current ML Initiatives: - Collaborating with data science team on predictive modeling projects, including logistic regression models for chronic homelessness classification to improve resource allocation for mental health services. ### Cloud Engineer & Technical Consultant @ Improving Jan 2019 – Jan 2023 | Alpharetta, Georgia, United States Cloud engineer and technical consultant for Improving's diverse range of clients. Projects include the following: 1) Catalis: Common Portal MVP • Migrated on-premises application to AWS and configured IAM, VPC, EC2, EBS, and S3 integrations focusing on HA. • Developed front-end components and services using Angular and TypeScript. • Increased Angular integration test and unit test code coverage by 80% via Jest. • Created CI/CD pipelines and CloudFormation templates to automate code and infrastructure deployments. 2) UCare: CORE Project • Designed and implemented healthcare SOA applications using C# and .NET, integrating into Angular frontend. • Enhanced API runtime efficiency by 32% by aggregating legacy data via SQL and PL/SQL stored procedures. • Implemented Step Functions to invoke long-running serverless workflows handling data transformation needs. 3) Donatos: Customer Data Migration • Implemented serverless RESTful APIs in API Gateway via Python and Django, decoupling application tiers while reducing integration response payload sizes by 75% and overall API latency by 50%. • Utilized SNS with SQS Fanout to distribute integration transformation jobs to AWS Lambda functions. • Migrated legacy customer data into a new normalized RDS Postgres database, reducing overall size by 90%. • Documented RESTful API using the OpenAPI specification to streamline developer onboarding. 4) Assurant: MCI4 Rehost Project (MyCoverageInfo v4.0) • Implemented Angular routing, dependency injection, and front-end components. • Developed backend RESTful APIs and services via C#, .NET Core, EF Core, SQL Server, SQS, and Lambda functions. • Increased policy retrieval speeds by 52% through an ElastiCache Redis implementation. ### Software Engineer @ LexisNexis Risk Solutions Jan 2018 – Jan 2019 | Alpharetta, Georgia, United States LexisNexis Risk Solutions: Motor Vehicle Report ETL Suite • Created an ETL suite in Python to pull data from SQL databases as well as various file formats, reducing manual data transformation efforts by 95%. • Developed Python Flask webapp UI and API to automate test data creation, accelerating release cycles by 62%. ### Software Engineer @ IndraSoft, Inc. An ECS Company Jan 2017 – Jan 2018 | Warner Robins Air Force Base, GA, United States 579 SMXS: DBM Application Migration and Rehost • Reverse-engineered a legacy Electronic Warfare database management program using Python, C#, .NET Framework, Entity Framework, and SQL Server. • Led requirements research, implemented backend business rules, and designed the updated GUI using PyQT. ### ARCNet Frontend and UX Developer @ Leidos Jan 2017 – Jan 2017 | Warner Robins Air Force Base, GA, United States ARCNet: UX Revamp & Offline CBT Development • Developed CBT (Computer-Based Training) programs for the Air Force Reserve Command's ARCNet application and updated squadrons’ ARCNet webpages to meet business and UX requirements. • Designed the UX for new CBTs and implemented RESTful APIs for completed CBT submissions. • Improved existing CBT development processes, reducing the time to completion by 400% from one month to one week. ## Education ### Master's degree in Data Science UC Berkeley School of Information ### Bachelor’s Degree in Information Technology Middle Georgia State University ## Contact & Social - LinkedIn: https://linkedin.com/in/colton-hester-660258106 --- Source: https://flows.cv/coltonhester JSON Resume: https://flows.cv/coltonhester/resume.json Last updated: 2026-04-10