Work
ProjectsCase studies, shipped work, and prototypes.Open SourceGitHub, npm, PyPI, and public packages.Hackathons & CompetitionsHackathons, contests, and rapid-build work.Depth
JourneyCareer timeline, roles, and education.AchievementsScholarships, rankings, and milestones.Courses & CertificationsStructured learning, courses, and credentials.SkillsTechnical capability map.Writing · 02
I write when a problem keeps returning: a slow query, an AI workflow nobody trusts, a codebase that fights its team, or a technical decision that deserves a more honest explanation.
Zero-downtime deploys on a cheap VPS
Field notes
495 published articles. Essays written here open in place; older tutorials take you back to the site where they first appeared.
Your website is missing from ChatGPT's search results. Learn why—and the five fixable reasons most sites get blocked, ignored, or buried.
TutorialA solid technical spec is your compass for AI agent projects. Learn what to include, how to handle uncertainty, and the template structure that cuts through ambiguity.
AI / MLProduction AI agents will fail. What separates safe systems from dangerous ones is detection, rollback, audit trails, and the protocols that restore control when things go wrong.
CareerMost technical portfolios fail because they show technology choices instead of problems solved. Here's how to build one that actually influences hiring decisions.
TutorialConnect your CRM to Slack with webhooks and event filtering. Automate notifications that drive action instead of creating noise.
AI / MLMCP is an open protocol that standardizes how AI applications connect to external tools and data. Here's why it matters.
ArticleMost organizations should budget 3-6 months for complete AI tool onboarding. Timelines depend on pilot scope, training depth, and how well you handle organizational change resistance.
AI / MLAI agents often feel sluggish—and it's not magic. Latency comes from concrete sources: sequential tool calls, wrong model choices, and context bloat. Here's how to diagnose and fix them.
TutorialA webhook is how applications talk to each other in real time. Instead of constantly checking for updates, one app instantly tells another when something happens—like a doorbell for the internet.
CareerMost teams that mentor through code review are actually just enforcing taste, inconsistently. Here's how to make review comments compound instead of repeat.
TutorialA monitor that checks the wrong thing, sends to a channel you don't watch, and pages you for every 30-second blip is worse than no monitor — it trains you to ignore it. Here's how to build alerting that actually works, for one person who eventually has to sleep.
CareerMost businesses don't need custom software. Here's how to tell if yours actually does—and why staying off-the-shelf is usually the smarter call.
AI / MLHuman-in-the-loop means keeping humans at critical decision points in AI systems. Learn when to automate, where to add oversight, and how to build accountable AI.
Cloud / DevOpsThis isn't a gate to keep you out. It's a readiness check — seven things worth being honest with yourself about before you're the one holding the pager, because a managed platform is still the right call for a lot of people right now.
TutorialCut no-show rates by 30-50% using automated reminders. Learn the timing, channels, and confirmation loops that actually work for service businesses.
MCP standardizes how AI agents connect to external tools and data, replacing per-model integration code with one protocol. Here's the architecture, the N×M problem it actually solves, and why it matters if you're building agentic systems.
TutorialMeta's Insights API powers automated reporting, but data consistency issues and authentication complexity break most first attempts. Here's how to build an agent that delivers accurate, trustworthy metrics.
CareerA structured discovery call separates automation projects that fit your expertise from those that don't. Learn the framework to uncover real problems, qualify scope, and set realistic expectations.
ArticleData pipelines automate the messy work of collecting, cleaning, and organizing business data. Here's what they actually do and why your business probably needs one.
FinTechA contracting rate that looks 40% higher than a salary isn't 40% higher take-home. Here's the actual math the IRS publishes, and the pieces that don't show up on either side's headline number.
FinTechWorking close to money forces habits most software teams can skip: real traceability, real correctness, real accountability for what breaks. Here's what carries over once you've built them.
FinTechThree instruments that all show up on an offer letter as "equity," taxed at three different times, under three different rules. The difference isn't fine print — it's when the IRS decides you owe money.
FinTechAn hourly rate isn't the number a salary converts to at 2,080 hours a year. It has to cover the hours nobody bills — sales calls, admin, the gap between contracts — plus a self-employment tax bill a salaried job never shows you.
FinTech"Six months of expenses" is a rule of thumb repeated so often it's stopped being examined. The number that actually matters isn't months — it's how volatile your income will be and how long clients take to pay.
FinTechLeave a day before your cliff and you own nothing. Leave a day after your grant is fully vested and a 90-day clock can still force you to pay cash for stock you may never see a return on, or lose it.
FinTechA competing offer moves a number. Being polite about it does not. Here's what the actual research on negotiation says works, and where the popular advice quietly stops being true.
FinTechNo invoice gets cut for a Saturday spent on a side project, so it feels free. It isn't. The hours have a market price, the IRS has an opinion on whether you can deduct anything against them, and unpaid maintenance has its own separate bill.
FinTechThe '4% rule' gets repeated as a guarantee. It's actually a backtest of US market history from 1926 onward, with named authors, a specific methodology, and limits both of them stated plainly. The popular version drops all three.
FinTechPublic companies have to disclose it in a 10-K when one customer accounts for a material share of revenue. A freelancer or a two-person studio has the same exposure and no filing requirement to force them to look at it.
FinTechA 0.3% offer and a 0.15% offer at the same company can be worth the same money, or double, depending entirely on a number the offer letter never mentions: the fully diluted share count it's a fraction of.
A working notebook
“Some ideas belong here because they connect directly to projects and services.
Older tutorials still live on Medium, where people first found them. This page keeps the trail intact without pretending everything was published in one place.
How I choose a topic
I would rather publish one useful explanation from lived work than five summaries of things everyone already knows.
Step 01
A topic usually starts as the second or third time the same engineering, product, hiring, or research question lands on my desk.
Step 02
I look for a shipped system, a measured result, a failed approach, or a public artifact that keeps the argument honest.
Step 03
Most technical choices touch something else. The useful links are the ones that help a reader follow that chain without opening fifteen tabs.
Step 04
Some pieces belong on this site; others already have a life elsewhere. I keep the original destination and publishing history intact.
Topics covered
The recurring technical and strategic themes represented across this archive.
A quick orientation
The practical details behind this mixed archive.
Those pieces were published there first. I would rather send you to the original article—with its date, responses, and history—than create a duplicate here.
Thirty-five posts are enough to become noisy. The filters separate original essays from outside tutorials, and permanent page links make it easier to pick up where you left off.
Usually. The essays here tend to connect a technical decision to the project, service, or research context behind it. External posts are often narrower, practical guides.
Please do. Real questions are where the best pieces start. Send the context, what you have already tried, and the part that still feels unclear.
Next step
You do not need a polished brief. A bottleneck, an uncertain architecture decision, or a product idea that refuses to become concrete is enough to start.