# Keith Mitchell > Principal Software Engineer in Boulder, Colorado. Senior Full-Stack / Platform Engineer — Enterprise SaaS, Analytics & Developer Platforms ## By the numbers - Years in software: 19+. Since April 2007, Leftcoastgeek.com co-founder. Professional web development since January 2009. - Years in enterprise SaaS: 11+. Rally Software, then CA Technologies, then Broadcom, October 2015 to present. Full-stack engineer with 11 years building and operating enterprise agile-management and value-stream analytics products at scale. Works across the entire stack and the entire lifecycle: React/TypeScript frontends, Python and Clojure backends, PostgreSQL analytics schemas, Kafka event pipelines, Helm/Kubernetes deployment, and Jenkins CI/CD. Keith must confirm hello@keithmitchell.dev and route it with Cloudflare Email Routing before it can be treated as monitored. The address is not taken from the resume. This site does not store messages. ## Pages - [Home](https://keithmitchell.dev/): Intro, experience comic strip, skill line, terminal, contact - [Printable resume](https://keithmitchell.dev/resume): Resume for print or PDF - [Resume Markdown](https://keithmitchell.dev/resume.md): Markdown twin of the resume - [Developers](https://keithmitchell.dev/developers): HTTP API and MCP documentation ## API - [GET https://keithmitchell.dev/api/v1/profile](https://keithmitchell.dev/api/v1/profile): Profile JSON - [GET https://keithmitchell.dev/api/v1/experience](https://keithmitchell.dev/api/v1/experience): Experience JSON - [OpenAPI](https://keithmitchell.dev/openapi.json): OpenAPI 3.1 description - [MCP](https://keithmitchell.dev/mcp): Streamable HTTP, stateless, read-only tools ## Optional - [llms-full.txt](https://keithmitchell.dev/llms-full.txt): Full resume text plus API notes ## Full resume markdown # Keith Mitchell Senior Full-Stack / Platform Engineer — Enterprise SaaS, Analytics & Developer Platforms Principal Software Engineer · Boulder, Colorado - Years in software: 19+. Since April 2007, Leftcoastgeek.com co-founder. Professional web development since January 2009. - Years in enterprise SaaS: 11+. Rally Software, then CA Technologies, then Broadcom, October 2015 to present. Email: hello@keithmitchell.dev Site: https://keithmitchell.dev Keith must confirm hello@keithmitchell.dev and route it with Cloudflare Email Routing before it can be treated as monitored. The address is not taken from the resume. This site does not store messages. ## GitHub - [https://github.com/leftcoastgeek](https://github.com/leftcoastgeek) - [https://github.com/leftmostgeek](https://github.com/leftmostgeek) ## Professional summary Full-stack engineer with 11 years building and operating enterprise agile-management and value-stream analytics products at scale. Works across the entire stack and the entire lifecycle: React/TypeScript frontends, Python and Clojure backends, PostgreSQL analytics schemas, Kafka event pipelines, Helm/Kubernetes deployment, and Jenkins CI/CD. Consistent track record of owning capabilities end to end — from first commit through security hardening, on-premises packaging, release automation, and coordinated multi-service rollout — rather than handing work across team boundaries. Recent focus: delivering a predictive outcome-intelligence product as both a SaaS surface and a self-contained on-premises install, and building LLM-facing API surfaces for agent-driven tooling. ## Experience ### Principal Software Engineer — Rally Software → CA Technologies → Broadcom 2015 – Present Boulder, Colorado Engineer on an enterprise agile-management and value-stream-management platform, spanning the flagship web application, the analytics backend, and a predictive outcome-intelligence product line. Sr. Software Engineer at Rally Software from October 2015. Principal Software Engineer at CA Technologies in Boulder from July 2016. Rally Software was acquired by CA Technologies in 2015, and CA Technologies by Broadcom in 2018. One continuous tenure. #### Outcome Intelligence · 2024 – Present Stack: React, TypeScript, Vite, MUI, Highcharts, React Router, react-intl, Vitest, Playwright, Mock Service Worker - Built a product analytics and outcome-forecasting web application from the ground up as primary frontend author — covering portfolio analytics, DORA and Flow metrics, outcome and goal modeling, and predictive health metrics. Delivered the main dashboard, a query builder, CSV export for measures and targets, and charting with ideal-threshold plot bands and business-rule overlays. - Shipped predictive goal-setting and forecasting features — a second-generation forecaster, an AutoML lag-indicator pipeline with loading and confidence states, outcome-prediction gauges comparing current against suggested trajectories, correlation display, and persisted optimizer bounds (custom constraints) that survive page reloads and URL changes. - Modeled hierarchical and chained outcomes end to end — parent/child linking with and without weightings, read-only treatment of inherited outcomes, drill-down navigation from an outcome summary into its chained children and back, and cascading deletes that clean up metric relationships. - Ported the entire frontend onto a new Python backend, re-keying metric and outcome identity across the codebase, replacing the router and link layer, adding a real 404 route, and removing dead feature-toggle and localization code in the process. - Rebuilt the dashboard information architecture — scoped the dashboard to a single outcome tree reflected in the URL, added a list view plus unified search across all view modes, delivered per-route contextual help anchored in-page, and gated development-only surfaces to non-production deployments. #### Value-stream automations backend · 2026 – Present Stack: Python, PostgreSQL, JSONata, OpenAPI/Swagger, PyInstaller, React/TypeScript, Docker, Helm, Kubernetes, Jenkins, Artifactory, pytest, JWT/JWKS, RHEL 8/9, Rocky Linux 8/9 - Led the port of a Clojure analytics backend to Python, replacing live third-party API reads with local PostgreSQL queries and reimplementing the automations engine — an HTTP request plus a JSONata transform, recorded as a measure against a metric — along with its encrypted secrets store and runtime credential injection. - Designed and shipped a secure cross-product launch flow — verification of a signed launch JWT against a remote JWKS endpoint, a durable session with a sliding activity-based TTL, and session transport hardened against XSS and CSRF, iterating between header- and cookie-based approaches (including CHIPS-partitioned __Host- cookies) based on threat analysis. - Packaged the product as a self-contained on-premises install — a PyInstaller build sourcing a hermetic interpreter, bundling frontend assets, API documentation UI, and all runtime resources. Built once natively on RHEL 8 and smoke-tested per target OS against a Helm-provisioned PostgreSQL, then published to an internal artifact repository. Validated across RHEL 8/9 and Rocky Linux 8/9. - Built the release-branch automation the team now uses to cut release and hotfix lines — an atomic script that cuts matching branches across the service and its deployment-chart repository, a versioning scheme so release builds never collide with mainline or overwrite the latest tag, a dedicated no-bump publish job, and a discoverable release runbook. - Hardened the service across its attack surface — externalized logging configuration and redacted a credential leak from log output, restored SSRF blocking in proxy-request handling, corrected multi-tenant secret scoping so private credentials resolve in previews without crossing tenant boundaries, fixed Content Security Policy to unblock inlined assets, patched a published dependency CVE, and hardened application settings against misspelled environment variables and cleartext secret exposure. - Enforced a product-wide naming boundary, preventing internal codenames and upstream vendor system names from reaching end users through UI text, API payloads, or run logs. - Improved CI throughput by parallelizing test and build stages, serializing mainline builds to make the version-bump push atomic rather than retry-prone, and routing dependency installation through an internal mirror. #### AI / LLM-facing platform work · 2026 Stack: Java, Gradle, Model Context Protocol (MCP), OpenAPI 3, Swagger UI, JSONata, Clojure, Python - Extended a Model Context Protocol server that exposes the platform's APIs as tools for LLM clients — adding an analytics-product toolset with authenticated OpenAPI spec fetch, correcting bearer-token handling (stopping a JWT from being forwarded as a session identifier to a downstream API), and gating the toolset on module availability so it activates only where licensed. - Made two backends consumable by AI agents — merged fragmented OpenAPI specifications into a single coherent document with meaningful tags and descriptions, added LLM documentation endpoints, authored a terse llms.txt guide for agent-driven API calls backed by real files rather than inlined strings, and served API documentation UI fully offline with working session-based authorization, eliminating a third-party CDN dependency. #### Frontend platform engineering · 2015 – 2019 Stack: React, Redux, JavaScript, SCSS, Webpack, Jest, i18n, Java/JSP, Node, Jenkins, Docker Largely pair-programmed, per the team's practice at the time. - Shipped a workspace-relative time-zone program across the full stack — localized date rendering in React, the server-side CSV renderer, and workspace-configuration handling, delivered behind a feature toggle and closed out by removing the toggle entirely. - Extracted a legacy JavaScript framework from a live React application, incrementally and behind a toggle — replacing its session-storage, notification, navigation and editor dependencies on the frontend while moving the framework into a bundle and rewriting server-side templates to match. - Rebuilt a legacy Kanban board as a new product surface — add-new behavior, quick filters, pagination with archive filtering, work-in-progress counts, and editor/viewer permission handling. - Built board swimlanes end to end — the data transform, row and header components, the grouping strategies (owner, parent, timebox), and the end-to-end specs. - Contributed early application-security work — sanitizing user-controlled board preferences before persistence and rendering, moving from a hand-rolled regex to a maintained HTML-stripping library. - Retired three legacy feature toggles and the code paths behind them, stabilizing roughly 93 end-to-end specs through the removal. - Owned shared-component release plumbing — the version bump script, its CI job, the Docker build node, and the golden-image job backing visual-diff regression testing. - Delivered a reusable fetch/Redux data component, a Webpack 2 upgrade, a Node 8 upgrade across three repositories, performance instrumentation wired into the e2e harness, and a codemod converting absolute imports to relative paths. #### Analytics and platform engineering · 2020 – 2025 Stack: React, Highcharts, Java, Groovy/Jenkins, Kubernetes, Helm, Docker, Apache Spark - Onboarded and decommissioned services across ephemeral Kubernetes test clusters — container images, manifests, cluster spin-up and redeploy scripts, CI job definitions, and nightly destroy-and-rebuild automation. - Built Spark-on-Kubernetes job plumbing and shared CI pipelines for a Clojure microservice estate. - Delivered program-increment cumulative flow diagrams and Portfolio Kanban charts, plus a reusable instrumented fetch API and client-metrics instrumentation. - Delivered a cross-team dependencies widget and the analytics dependencies and overview pages, and took a major analytics release to general availability — including retiring the feature toggles and legacy routes behind it. - Built a team distribution chart, a cycle-time chart handling partial workflow states, and a work item distribution chart, including removing its compliance-tier gating flags for general release. - Fixed defects across data tables and boards — permission-aware add-new button logic rolled out behind a toggle, swimlane collapse behavior, null group handling, and saved-views behavior — and contributed to an incremental class-to-functional-component migration. #### Analytics backend and ETL · 2022 – Present Stack: Clojure, Leiningen, deps.edn, Babashka, PostgreSQL, SQL, Liquibase, Apache Kafka, Avro, schema registry, Netty, JSONata, clj-kondo, cljfmt, Docker - Implemented change-tracking on the core flow-state fact table in the shared analytics library and the corresponding cumulative-flow pre-calculation in the ETL replay service, scoping computation to the relevant work-item flow states, and later optimized second-generation query performance. - Maintained and secured the analytics backend — added CSV and service-layer export routes, a timeout guard on untrusted transform parsing, secret scrubbing in logs, request-parameter identifier filtering, and a mail-library migration to Jakarta. - Remediated a coordinated security-defect batch across the analytics backend: OAuth login CSRF via state-nonce validation, OAuth client secrets moved from source to environment configuration, cross-workspace read prevention by scoping job logs to a validated definition, workspace scoping on a list endpoint, a request-body allowlist blocking cross-tenant secret overwrite, SSRF fixes, and relocation of a webhook credential from source control into managed CI credentials. - Fixed security findings from static analysis in the analytics frontend — stored HTML injection in chart titles, tooltips, and series names; unsanitized upload-error HTML; and a one-year cache TTL incorrectly applied to mutable deploy artifacts. #### Platform, CI/CD, and infrastructure · 2022 – Present Stack: Jenkins (declarative pipelines and Job DSL), Groovy, Docker, Kubernetes, Helm, Ansible, Artifactory, Apache Kafka (mTLS, truststores, CA certificates), PostgreSQL, Bash, Maven - Drove a multi-service Kafka mTLS rollout across seven repositories — two backend services, their deployment charts, the shared Kafka client library, production environment values, and ephemeral test-cluster configuration. Made broker CA trust work without a client certificate for on-premises deployments, reused the existing CA secret rather than minting a new one, and changed startup to fail loudly instead of silently dropping message publishes. - Executed a full service decommission across six repositories — removing the service from the shared CI job definitions, the deployment-chart dependency graph, the test-cluster values template and cluster spin-up/redeploy scripts, database administration tooling, and the shared PostgreSQL role and schema baseline. - Eliminated external dependency-resolution risk across a JVM service estate — migrated projects off public package repositories onto an internal mirror, resolving recurring rate-limit build failures, distributed shared build configuration, and modernized one project's dependency tooling. - Closed a supply-chain vulnerability by adding checksum verification to a vendored third-party download script and pinning a CI script to a verified commit SHA; migrated formatter distribution to a single-source internal helper with a backward-compatible fallback so existing pipelines kept working throughout the transition. - Authored and maintained shared CI for a microservice estate (Groovy Job DSL) — including manual release-branch and branch-agnostic publish jobs, and migrating a frontend deploy pipeline from object-storage upload to a container-image deploy. - Authored Kubernetes deployment charts for multiple services — exposing logging configuration, synchronizing environment variables with application settings, mounting Kafka mTLS truststores, and replacing single-URL configuration with structured multi-tenant configuration. - Sustained dependency and CVE hygiene across roughly fifteen services — transitive dependency pin alignment, security upgrades to serialization and sanitization libraries, logging-framework remediation, SCA finding resolution, and credential rotation. #### Engineering practice Stack: Playwright, Vitest, Mock Service Worker, pytest, testcontainers, Jest, clojure.test - Built out automated test infrastructure for the analytics frontend — end-to-end tests in CI with recorded flows, browser-mode unit testing, service-worker-based API mocking with a development toggle so the application runs with no backend at all, local TLS certificates for the mock worker, and type-level tests. - Established agent-assisted development workflows — repository-level AI coding configuration, automated code-review and security-scan tooling, and onboarding documentation across several repositories. - Consistently reduced surface area while adding features — paired feature work with refactors that delete dead code and shrink diffs: removing convenience pages and everything only they kept alive, deriving UI state rather than synchronizing it through effects, and replacing inlined string content with real files. ### Senior Software Engineer — Kapost Mar 2015 – Sep 2015 Greater Denver Area ### Senior Software Engineer — Quick Left Apr 2014 – Mar 2015 ### Development Lead — Datu Health Jun 2013 – Apr 2014 Boulder, CO ### Engineer — Quick Left Jan 2012 – Jun 2013 Boulder ### Web Developer — Markit Jun 2011 – Jan 2012 Boulder, Colorado ### Web Developer — Wall Street On Demand Jan 2009 – Jun 2011 Boulder, Colorado ### Co-Founder — Leftcoastgeek.com Apr 2007 – Aug 2008 ## Education - Geological Engineering, University of Arizona, 2000–2005 ## Technical skills - **Languages** — TypeScript, JavaScript, Python, Clojure, Java, Groovy, Bash, SQL (PostgreSQL) - **Frontend** — React, Redux, Vite, MUI, Highcharts, React Router, react-intl (i18n), Vitest, Playwright, Mock Service Worker, ESLint/Prettier - **Backend** — Python (ASGI/OpenAPI services, PyInstaller packaging), Clojure (Leiningen, deps.edn, Ring/Compojure-style HTTP services), Java (Gradle), JSONata transforms, OpenAPI/Swagger, Model Context Protocol (MCP) - **Data & Messaging** — PostgreSQL (analytics schemas, Liquibase migrations, ETL and pre-calculation pipelines), Apache Kafka (producers/consumers, Avro, schema registry, mTLS), Cumulative-flow and flow-metric computation, Time-series forecasting surfaces - **Infrastructure & CI/CD** — Jenkins (declarative pipelines and Job DSL in Groovy), Docker, Kubernetes, Helm (chart authoring and unit testing), Ansible, Artifactory, Ephemeral test clusters, RHEL/Rocky Linux - **Security** — XSS and HTML-injection remediation, SSRF controls, OAuth CSRF/state-nonce validation, Content Security Policy authoring, Secure cookie design (CHIPS-partitioned __Host- cookies), JWT/JWKS verification, Multi-tenant data scoping, Secret encryption and log redaction, Supply-chain hardening (artifact checksum verification, dependency pinning), CVE and SCA finding remediation - **AI tooling** — MCP server development, LLM-facing API documentation (llms.txt, OpenAPI for agents), Agent-assisted development workflows and automated code review Last updated: 2026-11-18 ## MCP tools - get_profile: name, title, location, email, years in software, years in enterprise SaaS, GitHub sameAs, summary, knowsAbout, worksFor, alumniOf - get_experience: tenure, focus areas, earlier roles, education, and skill groups - search_skills: case-insensitive search of the Technical Skills section - get_contact: email, location, and GitHub URLs Every tool is read-only. Input objects reject unknown keys.