SD-01: Support Desk
Agents propose, code decides: a policy guard, a QA critic and an executor that re-checks its own limits.
Computer science degree, three years of hardware-in-the-loop testing at Nokia, and prototypes built like production code. Move across the screen to see the system under the product.
Agents propose, code decides: a policy guard, a QA critic and an executor that re-checks its own limits.
3 builds
Three independent builds. Each has working code, the documents a team would need, and the numbers from its own test runs.
An adaptive running coach, run as a full program: from charter and RAID log to a tested FastAPI service and an iOS companion. Rules make the call; the LLM only explains it.
Wearable ingestion, SQLite persistence, Docker, CI, a groundedness checker and a fallback when the LLM fails, covered by 171 tests.
| Automated tests | 171 |
| Architecture decision records | 5 |
| Tracked risks, with triggers | 7 |
| Critical-path dependencies | 5 |
| Baseline safety violations | 0 |
Agents that resolve support tickets end to end, refunds included. The model can only propose an action; plain code decides whether it happens.
Idempotent executor, defense in depth (the executor re-checks limits itself), bounded revision loop, per-step traces.
| Eval tickets passing | 19 / 19 |
| Wrong or unauthorized actions | 0 |
| Promises that didn't happen | 0 |
| Must-escalate reaching a person | 100% |
| Resolved without a person | 63% |
Offline runs use a stand-in model with injected faults. They test the safety net, not live-model quality.
A data product that warns a consignment retailer when contracts are falling behind target, while there's still time to act. I pre-registered the test, and it said my first bet didn't hold.
BigQuery views, a seeded data generator, data-quality checks and a backlog synced to GitHub Issues.
| False alarms, first bet | 80% |
| False alarms, after the fix | 36% |
| Warning before contract ends | 30 days, vs 0 today |
| Question layer, Sonnet | 30/30 at $1.08 |
| Question layer, Haiku | 29/30 at $0.05 |
Fictional company, simulated data of 1,968 contracts.
Most recent first.
AmazonJun to Aug 2026
Etna, Ohio
Thomson ReutersNov 2023 to Jul 2025
Bangalore, India
NokiaSep 2020 to Oct 2023
Bangalore, India
PhilipsJan to Jun 2020
Bangalore, India
How I think on paper: working-backwards docs, roadmap trade-offs, case answers, and teardowns.
What to build, in what order, and why, stress-tested against delays and alternative schedules.
Pricing a two-sided student storage marketplace from survey data. Answer: yes, narrowly, at about $55 a month.
Tailors a résumé to a posting without inventing experience. A code-based critic blocks new numbers and inflated verbs.
Solo product and design: onboarding, custom tinting on one glanceable canvas, and forgiving refill flows.
Twelve weekly teardowns, all shipped on time: onboarding, recommendations, notifications and payments.
I came to product through a computer science degree, three years of hardware-in-the-loop quality work at Nokia, and a product seat at Reuters where I learned what it costs when a trading desk's SLA slips. This summer I worked on fulfillment operations at Amazon. Now I'm at USC Marshall as a Dean's Merit Scholar, sharpening the strategy and finance side.
I work best where the brief is a little unclear and the engineering is a little hard. Outside work I founded Trektroops, which grew from a 12-person group to 570+ paying members, held together by a leader-rating system that kept quality up across 50+ trips at once.
I'll happily talk cricket analytics, sim racing, personal finance or espresso.