Ticket
Request, issue, PR, or captured failure
Ship AI capability through provider abstraction, eval gates, tenant boundaries, and auditable output.
Work is grouped by shipped system, not by resume chronology.
Provider choice is an implementation detail. The operating system is the contract: route the work, choose the model, set the effort, verify the evidence, and hand off with scope intact.
Request, issue, PR, or captured failure
Select skill/playbook from context
Load rules, constraints, and verification paths
Choose provider/model without locking feature code
Set reasoning budget by risk and ambiguity
Isolated branch-local execution
Focused tests, lint, build, and screenshots
Proof, notes, and decision record
Review-ready state with exact scope
AGENTS.md, compatible rule files, reviewer roles, local hooks, preflight scripts, focused verification, and process notes keep every contributor operating inside the same boundaries.
SuperOrgs development was run through a custom Claude Code workflow stack: tickets became isolated worktrees, agent skills executed the loop, subagents reviewed the work, browser QA verified behavior, and hooks failed closed before handoff.
/sup-autoLinear ticket execution from brainstorm and plan through TDD, implementation, multi-agent review, browser verification, and PR handoff.
/ticketEnvironment setup, scoped execution rules, and codified guardrails before implementation begins.
/superorgsQ&A over the codebase, Linear, GitHub, and project context for faster orientation.
/autopilotOrchestrates worker skills into one operating cycle instead of one-off prompts.
/babysit-prsTracks CI, behind-main drift, review aging, and PR readiness.
/sup-queueFeeds queued tickets into the autonomous execution workflow.
/deps-bundlePartitions dependency updates into safe batches versus held upgrades.
/skill-retroTurns friction logs into concrete skill and process improvements.
/shipMoves implementation through verification, review, and handoff as one delivery path.
Click a cluster to highlight the systems that used it.
Compact by default. Expand a row for the evidence behind the role.
Built a Laravel data warehouse for HR workflows, generated historical data for 1,000+ employees across more than 10 years, and modeled reviews, increments, salaries, allowances, clients, and managers.
Sep 2018 - Jan 2019Co-authored the FIRE'18 MAPonSMS author-profiling work on multilingual Roman-Urdu and English SMS text, using stylometric, Bag-of-Words, and psycholinguistic features to predict author gender and age.
CEUR-WS / FIRE 2018Created a Ruby API wrapper for Oracle Eloqua application APIs covering campaign operations and Eloqua asset types in the initial version.
Ruby gemBuilt web/mobile mockup generation work for low-code/no-code application design workflows.
React / Node.jsIf the work needs AI systems that survive real product, tenant, security, and developer-workflow constraints, this is the lane.