Ryan Rampersad

2024 - History


October

  • Built a prototype for rudimentary alternative to data warehouse style reporting
    • Hypothesized that "sql-as-the-interface" was critical for reporting purposes
    • Proposed that existing integrating routes were insufficient
    • Explored an alterative method: adhoc creation of SQLite databases mirroring the schema and select rows from the parent data set
    • Shared findings with broader team and discussed ramifications of the approach
    • Used this as an opportunity to use ChatGPT Enterprise code generation
  • Discussing secrets management approaches for various vendor services operated through Terraform
    • Internal enterprise accessibility (UI)
    • Programmatic access in Vela (TF) and as a Workload (API)
    • Low operational overhead
    • High coherence with existing tooling
  • Updated vulnerability reporting to use v2 capabilities
  • Fixed bugs as they arose from the new v2 capabilities
  • Prototyped a new internal repository SVG badge for easier at-a-glance-status reports
    • SVG badge using basic go templating
    • Data sourced from cloud systems on uptime, deployment status, etc

September

  • Continued evaluating various SCA and SAST vendors
    • Increasingly scrutinizing vendor capabilities and weighing pros and cons
    • Supporting critical decisions with detailed reports
    • Compiled results of at least five SCA and SAST vendors
  • Continued as primary contributor to the cloud resources user interface codebase
  • Helped onboard new team members so they have at least a fundamental understanding of the codebase
  • Contributed to platform capabilities
    • Based on a feature request, contributed a basic integration with the API Gateway to determine if a policy was active or inactive, and report that during API deployment accordingly
    • Followed existing logic and codebase patterns
    • Wrote tests and increased Sonarqube coverage by 1.5%!
  • Programmatically alerted teams to stale deployments
    • Opened issues with details on their deployments in their respective repositories
    • Shared important details regarding the stale deployment cleanup timeline
  • Performed maintenance on various deployed applications
    • Upgraded dependencies
    • Resolved existing vulnerabilities
  • Continued contributing and brainstorming on new tech strategy

August

  • Continued evaluating various SCA and SAST vendors
    • Planned replacement scenarios internally
  • Began working on a new technical strategy initiative
    • Helping with various aspects
  • Continued as primary contributor to the cloud resources user interface codebase
  • Went to the Minnesota State Fair three times! 🎉

July

  • Began testing various SCA and SAST vendors
    • Evaluating findings reliability
    • Evaluating integrating capabilities (out of the box, API, CLI, etc)
    • Developed baseline SCA and SAST projects in JS/Node and Java/Spring Boot
  • Assisted the CIAM team with the Okta Identity Engine upgrade
    • Helped review bugs as they were filed
    • Created prototypes of different scenarios so product teams could follow along with necessary upgrades
  • Documented the legacy vela plugin deprecation timeline and approach
    • Explicitly listed existing plugins that are being deprecated
    • Created THEN-vs-NOW style documentation to help teams gradually migrate their pipelines
    • Wrote about the impetus and motivation behind the deprecation
    • Planned new code instrumentation, beyond metrics gathering, to gradually slow and eventually break builds if legacy plugins are still used
  • Continued as primary contributor to the cloud resources user interface codebase
  • Saved an old project as a memorialized frozen version and kept it for future generations

June

  • Monitored AEM usage until official shutdown
    • Finished data and image migrations
    • Finished codebase updates where necessary
    • Repeatedly communicated shutdown dates
    • Celebrated wild success for our years-long cutover
  • Became a primary contributor to the cloud resources user interface codebase
    • Implemented react query and react table with appropriate abstractions to ensure flexibility while offering some reusability
    • Guided team to use an openapi to sdk generator to make integrations easier
    • Enforced minimal customization to MUI to ensure future upgrades were smooth
  • Participated in a pilot program, field testing ChatGPT Enterprise
    • Framed around "engineering capabilities"
    • Set a basic cadence for reviewing findings and experiences (positive and negative) with the tooling
    • Shared video demos with the broader group regularly

May

  • Created end-to-end automated scripts that saved all existing AEM-Scene7 images locally, and programmatically reuploaded them in the new public-image target, Cloudflare Images
    • Created a data tracking system with caching to efficiently upload images
    • Updated various infrastructure pieces (CMS, S3-abstraction) to handle new workflows
    • Fixed various bugs introduced by changing internal vendor support
    • Wrote extensive documentation to guide product teams on minimal necessary changes required in their application layers (primarily limited to URL swaps)
  • Continued to monitor AEM usage
    • Created and watched datadog dashboards that summarized legacy ecosystem paths typically associated with AEM
    • Notified teams of remaining as shutdown looms
  • Rewrote daily productivity booster "gclone"
    • Written in plain bash
    • Intelligently clones a repository into a known folder path based on org/repo name
    • Automatically cd into the repo after clone
  • Wrote new tutorials for first time contributors
    • Wrote new Python API tutorial featuring FastAPI and Docker and Vela
    • Wrote new Java Spring Boot API tutorial also featuring Docker and Vela
    • Extensively updated existing guides for SPAs using Vite
  • Began assisting another platform team focused on cloud operations with their custom ui offering
    • Helped coach a backend engineer on frontend tooling and patterns
    • Helped scaffold the initial application using MUI, Vite, TypeScript, React (partially based on previous work)
  • During an innovation hackathon, used Ollama + Llama2 and Llama3 along with the GitHub API to pull activity for repositories (PRs, commits, change sets, etc) and summarize daily/weekly activity in a concise format
  • Created the CIAM organization while assisting Okta IDP team migration efforts

April

  • Continued Adobe Experience Manager shutdown evangelism
  • Ran two brownout periods, stage and then production, incrementally to help teams safely discover lingering integrations with legacy tooling
  • Worked with various teams to ensure a smooth transition
    • Helped a team using a legacy flow to onboard to the current reverse proxy solution
    • Helped teams previously using Apache migrate to containerized nginx for server side routing
  • Created a suite of basic single purpose vela plugins
    • trigger-deployment-plugin
    • trigger-release-plugin
    • validate-tag-plugin

March

  • Began preparing for final Adobe Experience Manager shutdown
    • Published documentation, featuring timeline, explanations, brownout dates, diagrams
  • Swapped internal Sonarqube instance authentication workflow from SAML to OIDC
    • Swap decreased workflow time substantially
  • Instrumentation legacy internal vela plugins with DataDog event hooks
    • Instrumentation provided metrics to prioritize updates and deprecation of legacy internal vela plugins
  • Created "undisclosed-ui" as a demonstration of a reproducible frontend codebase
    • Featuring Vite, TypeScript, MUI, TanStack tools
  • Restricted "net new" deployments in a legacy organization to spur product team scoped application management as previously planned
    • Allowed existing deployments to continue to operate with no firm migration timeline out of the legacy organization

February

  • Triaged unexpected behavior between vela and github
    • Tag events were no longer associated to their reference branch when tagged from the pipeline itself in undefined conditions
    • Created reproduction cases
    • Created an workaround that used the preferred interface anyway—using the github release ui
  • Experimented with an Okta Auth POC to extend token duration
    • Okta access tokens expiration is extensible while id tokens expiration are not
  • Refined a public authentication workflow for external and internal users
    • Previously using a non-OIDC approach with old versions of authentication packages
    • Previously implemented with a cookie-ful approach in mind; and later surreptitiously updated to use session tokens
    • Entirely re-implemented the flows using latest generation libraries
    • Reduced the number of SFDSCs (Screen Flash During State Change) and redirects and improved overall user cognitive load
  • Triaged a major security incident
    • Created a reproduction
    • Created initial documentation
  • During a hackathon, created a GitOps flavored link shortener
    • Used hono, zod, yaml

January

  • Announce that a monolithic cloud cluster will be locked for net new deployments broadly
    • Provided documentation for this plan
    • Applied a 4-sprint—60-day window to give teams ample time
    • Provided value aspects for breaking breaking this cluster a part if important and strategic
  • Finalized and began implementing with the api for arbitrary datadog events
    • Began refactoring and adding the curl -X POST call to the api service in various vela plugins to begin getting telemetry
    • Created a datadog dashboard with sections scoped to each plugin's usage
  • Backed up legacy application s3 buckets using various scripts
  • Backed up legacy application Auth0 databases
  • Compiled a list of AWS and Cloud stakeholders to get explicit documented approval to shutdown legacy application accounts
  • Shutdown legacy application accounts and documented their official closure dates
  • Enjoyed a wonderful winter break vacation for 10 days

Browse by Year


513 weeks