Joseph Krall

Principal Software Engineer · Practical AI in Engineering · PhD, Computer Science

Washoe Valley, NVkralljoe@gmail.comjosephkrall.comlinkedin.com/in/joseph-krall

Summary

Hands-on, full-stack principal engineer with a PhD in Computer Science and 16 years of building software: NASA flight research, energy analytics, industrial machinery monitoring, and mortgage and title production platforms. I wear whatever hat the work needs: take fuzzy requirements from product, turn them into architecture and user stories, lead the team through them, and build many of them myself. My edge is AI: I work with it like a keyboard, every day, without ceremony, and I build the rules and reviewers that keep it honest.

Experience

Cloudvirga (part of Stewart)

Principal Software Engineer · Jun 2021 - present · Remote

Title Production (current)

• Bridge product and engineering through Stewart’s refocus on title production: pin down requirements with product in discovery, write the features and user stories, lead grooming, then architect and build alongside engineers and QA. Keep the cross-functional team near 50 points a sprint amid real ambiguity.
• Spearheading ResWare integration through a shared Stewart integration service, so our title platform reuses one path instead of building its own.
• Leading an AI chatbot for buyers’ FAQs: team prototypes, lessons from product leaders, cost and risk controls (rolling summarizer, prompt caps, prompt-injection guardrails, LLM trials), and a planned FAQ and workflow knowledge base retrieved only when needed (RAG). Personal data is deidentified.
• Architected the buyer messaging center and task-based forms: admins customize questions, title officers send forms to buyers, and “message us” from a task carries that task’s context to the title officer.

Across Engineering

• As tech lead, raised sprint velocity on two teams from about 30 to 50 points and from about 40 to 60, and improved DORA metrics alongside that.
• Built an AI pull-request reviewer (several LLMs under the hood) that engineers choose on about half of all PRs, even with GitHub Copilot’s review one click away. Wrote the context rules that let AI coding agents work reliably in our large .NET codebase.
• Authored engineering standards (AI governance, API design, .NET coding, UX) and the .NET style guide: evidence over taste, and leading by example so teams adopted them without being ordered to.
• Created MIND UP, a company course on generative AI at work: safety, ethics, prompt craft, and a “try it first” habit of doing your own thinking before AI review.
• Designed our engineering interview challenges, from a Zelda-rupee puzzle (count to 999 in 8-bit storage) to questions on code review and leadership, and evaluated 15+ candidates.
• Ran Orchard (organizational risk discovery), biweekly sessions with the engineering director that surfaced unowned risks across product, engineering, sales, and marketing before they became problems.

Mortgage Platform

• Led the company-wide move of our Encompass integration from its legacy SDK to the modern API, a 10-15x performance gain: average loans from about 30 seconds to 3, large loans from 2.5 minutes to about 10 seconds. Wrote a custom queue for the new API and built a strong working relationship with the vendor’s key engineer.
• Built on a distributed, message-driven microservices platform: Azure Service Bus queues and topics, RabbitMQ in legacy services, Redis, and Docker and Kubernetes in the build pipeline. Our production clients ran on AWS.
• Partnered with DevOps on reliability: Application Insights and Splunk, dashboards and alerts on API latency, RAM, and CPU, and on-call war rooms that diagnosed outages fast and presented a unified answer to customers.
• Designed the Git branching strategy, led the TFS-to-Azure DevOps migration, and helped turn several waterfall teams into consistent agile ones.

Bently Nevada (GE / Baker Hughes)

Senior Software Development Engineer in Test (SDET) · Feb 2017 - Jun 2021 · Minden, NV

• Built security-critical machinery condition monitoring software used by industrial customers worldwide.
• Wore both hats: wrote the product code and the automated end-to-end tests that guarded it.
• Mentored four junior engineers, led agile planning and design sessions, and interviewed candidates.

LoadIQ

Chief Data Scientist · Oct 2014 - Feb 2017 · Reno, NV

• Built unsupervised algorithms that split a building’s single power signal into individual appliances, using graph theory over steady states (Python, NetworkX, scikit-learn).
• Ranked second in EPRI’s 2016 independent evaluation of energy disaggregation technologies.
• Presented Graphical Closure Rules for unsupervised load classification at the NILM 2016 workshop.

West Virginia University

Doctoral Research Engineer · Aug 2010 - Aug 2014 · Morgantown, WV

• Created GALE, a multi-objective optimizer that finds good solutions in 30-50 model runs where standard genetic algorithms need thousands (IEEE Transactions on Software Engineering, 2015).
• Built JMOO, an open-source Python framework for comparing optimizers fairly.
• Published in two IEEE Transactions journals (Software Engineering; Human-Machine Systems) and at AAAI.

NASA Ames Research Center

Research Scholar · Jun 2013 - Jul 2013 · Mountain View, CA

• Applied GALE to a simulation of pilots coordinating tasks on approach. Each model run took 7 seconds, so standard optimizers needed hours; GALE finished in minutes and found mitigations for missed or delayed tasks (IEEE Transactions on Human-Machine Systems, 2016).

Writing

Essays on Philosophy and AI

Author · 2026 - present · josephkrall.com

• Essays on working with AI: why it fills your blanks with the average of the web (Tabula Rasa), why it tells you what you want to hear (The Flatterer), and how to keep your judgment and voice when it does more of the work.

Education

• Ph.D., Computer Science · West Virginia University · 2014
• M.S., Computer Science · West Virginia University · 2010
• B.S., Computer Science & Mathematics · University of Pittsburgh at Johnstown · 2008

Recognition

• HousingWire Rising Star, 2024

Skills

• Languages: C#, Python, TypeScript, JavaScript, SQL, Java, C++
• Frameworks: .NET, WPF/XAML, Angular, React, Vue 3, Django
• Data: SQL Server, MongoDB, Azure Cosmos DB, PostgreSQL, MySQL, SQLite, Redis, NumPy, SciPy, scikit-learn
• Cloud & architecture: Azure (Service Bus, Cosmos DB, DevOps), AWS (working exposure), microservices, event-driven messaging, RabbitMQ, Docker, Kubernetes, REST APIs, distributed systems
• Observability: Application Insights, Splunk, dashboards and alerts, incident war rooms
• AI: Cursor, ChatGPT, Gemini, Perplexity; coding agents, context rules, LLM review and model selection, RAG, prompt-injection defense, multi-objective optimization
• Delivery: Git, CI/CD, automated end-to-end testing, Cloudflare Workers

Publications and projects: josephkrall.com/about · josephkrall.com/portfolio