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.
CareerCaffeine works by blocking your perception of a deficit, not by removing it. Combine that with a study showing sleep-restricted people can't feel their own impairment, and you get a mechanism worth understanding rather than a lecture about coffee.
Career10,000 steps was a 1960s pedometer marketing figure, not a finding. The research that does exist points somewhere lower and more useful — and it has something specific to say about whether you can type while walking.
CareerBIFMA certification tests whether a chair survives years of abuse, not whether it fits your spine. A handful of adjustability features have real evidence behind them; mesh versus fabric and "ergonomic" branding mostly don't. Price tracks the wrong list.
CareerThe controlled studies are real, replicated, and worth knowing — but they dehydrated people with exercise and heat, not a skipped water bottle at a desk. What they found lines up more with mood and fatigue than with raw reasoning ability.
CareerPhysiotherapists have spent decades telling people to sit up straight. The researchers who study this for a living went looking for evidence that a specific posture prevents back pain, and mostly came back empty-handed.
CareerThe popular version treats a bent neck as a mechanical injury waiting to happen. The systematic reviews tell a messier story: real associations, small effect sizes, and psychosocial factors that predict pain at least as well as posture does.
CareerA three-day weekend feels like it should undo a month of overtime. The sleep-debt research says it mostly doesn't, and the occupational-health literature on sustained long hours is blunter than most engineering teams want to hear.
CareerThe comparison is quotable and it's wrong by an order of magnitude. Smoking roughly triples all-cause mortality risk with no dose of exercise that cancels it out. Sitting's mortality association is real, smaller, and largely erased by an hour of daily activity.
Career"Sugar crash" is doing a lot of work it hasn't earned. Part of the 2-4pm slump is your circadian rhythm running on schedule with or without food, and only part of it is what you actually ate.
CareerMost posture advice is "sit up straight," which is unfalsifiable and useless. The elbow angle, monitor height and viewing distance parts are not — they're measured in degrees and centimeters, and the studies behind them are worth reading.
CareerGetting paged at 3am and simply sleeping two fewer hours are not the same injury. The research on interrupted sleep says the fragmentation itself, independent of total hours, is doing damage most on-call engineers never account for.
CareerA hard workout dips your executive function for a window afterward, then boosts it for a longer one. The research is fairly consistent on that shape and much messier on whether morning, midday or evening timing matters more than intensity does.
CareerWalking helps some kinds of thinking and measurably hurts others. The research is more specific than "movement is good" — it says exactly which tasks get worse, by how much, and which ones don't move at all.
CareerA randomized experiment at a Chinese travel agency found working from home raised performance 13% and halved attrition — and cut the promotion rate almost in half too. The same company's later hybrid trial, two remote days a week instead of five, found no such cost at all. The difference between those two results is the actual story.
CareerThere is no controlled trial settling home office versus coworking versus cafe for deep work, and this post won't pretend otherwise. What exists instead are well-established mechanisms — interruption-resumption cost, noise variance, coworking's real price, and the incidental-contact gap remote work opens up — that combine into an answer more useful than a single number would be.
CareerA single clarifying question with no context attached, sent to someone eight timezones away, can cost a full working day before either side notices. A 61,000-person study of Microsoft's move to remote work found the same failure mode at scale: fewer new collaborators, thinner cross-team bridges. Async communication is a trainable skill that closes exactly that gap. Most engineers never train it.
CareerOokla's Speedtest Global Index put the global median fixed-broadband speed at 102.48 Mbps in May 2025. Singapore sits above 400 Mbps. Cuba sits under 4. That's not a rounding difference for a remote engineer — it's the gap between a standup that works and one that doesn't.
CareerA study of over 61,000 Microsoft employees found that when synchronous overlap disappeared company-wide, collaboration across teams fell about 25% and the org chart got measurably more siloed — even though meeting hours barely moved. The real constraint on a distributed team was never raw hours of overlap. It's which decisions need a live room and which don't.
CareerIn Bloom, Liang, Roberts and Ying's randomized Ctrip experiment, home-based call center workers performed 13% better than the office control group and still got promoted at barely half the rate. Performance wasn't the bottleneck. Being seen was.
Career23% of roughly 3,000 remote workers picked loneliness as a top struggle in Buffer's 2023 survey — self-reported, not measured, but real. A separate study of 61,000 Microsoft employees found the more precise mechanism: cross-group collaboration fell 25% under remote work while scheduled meeting hours barely moved. The isolation is real. The fix most teams reach for, more scheduled calls, is aimed at the wrong half of it.
CareerA payroll system built to file one country's tax forms does not bend gracefully for a second country's address, and it usually fails silently — a direct deposit that quietly stops, not an error message that explains why. That is the actual failure mode of working remotely from abroad: rarely a dramatic visa problem, mostly a stack of small administrative systems that assumed you would never leave.
CareerA randomized experiment at a 16,000-person Chinese travel agency found working from home raised performance 13% and halved attrition — and also halved the promotion rate, conditional on that better performance. The difference between a company that fixes that and one that doesn't isn't a perk. It's where decisions get recorded.
CareerEvery photo of someone working from a beach is selling the one hour of the day that photographs well. The other twenty-three hours are timezone drift against a team that didn't move, a connection you can't test before you book it, and the gap between a two-week trip and a three-month one that almost nobody plans for separately.
CareerA 120-person randomized trial found no therapeutic effect on eye strain, regardless of whether participants were told the glasses would help. Cochrane's pooled review of 619 people agrees. If you already own a pair and like them, there's no evidence they're harming you either — this is about what the specific marketing claim can support.
Setting up separate dev, UAT, and production environments for a React Native iOS app in Xcode, without the usual scheme sprawl.
CareerA Cochrane review of 17 randomized trials and 619 people found blue-light filtering lenses probably make no difference to eye strain, vision, or sleep. The American Academy of Ophthalmology says screens don't emit harmful blue light levels. This is one of the more settled questions in eye health, and the popular belief is still wrong.
CareerIt was invented as a soundbite for media interviews, not derived from a clinical trial. No peer-reviewed study tested it directly until 2023 — a single 29-person trial with no control group. The rule is probably still good advice. The certainty it's usually delivered with is not earned.
CareerNot damage, mostly. A 1993 blink-rate study found screen use cuts blinking to roughly a third of its resting rate, and a 2023 meta-analysis of 66,577 people put symptomatic eye strain at 69% of screen users. Neither number means your eyes are being harmed.
CareerIt isn't one condition. "Digital eye strain" bundles ocular symptoms (dryness, burning, blurred vision) with musculoskeletal ones (neck and shoulder pain) under a single label, and different studies count entirely different things as qualifying for it.
CareerThere's no randomized trial pinning down an ideal brightness percentage — this is one of the few claims in this cluster where the honest answer is 'no rigorous study exists,' not 'the study says something different than expected.' What does have real backing is the principle behind the advice: match screen brightness to the room around it.
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.