Engineering essentials
The workflows we lead with: shipping, debugging, refactoring, reviewing, onboarding.
49 entries
Brutally critique my startup idea
Get the objections a skeptical investor would raise, before you spend six months finding them yourself.
Startups
Audit a design for accessibility before it ships, not after a complaint
Accessibility caught in review costs a design pass. Caught in production costs a support ticket and a redo.
Design
Build a fundraising narrative investors actually remember after the meeting
A deck of correct facts isn't a narrative. Find the one thing they'll repeat to their partners.
Startups
Debug the bug that only happens sometimes
WorkflowTurn an intermittent failure into a deterministic repro, then a named mechanism, then a one-line fix.
Coding
Critique a UI mockup like a design lead, not a fan
Get the feedback that finds the real problem, not a list of things you could tweak forever.
Design
Explain this error properly
Understand what actually broke and why, instead of pasting the first fix that compiles.
Coding
Build a cohesive visual identity from a blank page
WorkflowPicking colors and fonts before deciding what the brand needs to feel like produces a mood board, not an identity.
Design
Get productive in an unfamiliar codebase in one day
WorkflowBuild a working mental model of a repo you've never seen, and leave behind a CLAUDE.md that compounds.
Coding
Refactor across 40 files without breaking production
WorkflowPin the behavior first, change the structure second, and never let the two mix in one commit.
Coding
Review a pull request like a staff engineer
WorkflowFind the defects that matter, skip the ones that don't, and leave comments people act on.
Coding
Ship a feature with Claude Code without rewriting it twice
WorkflowThe explore → plan → slice → adversarial-review loop that senior engineers actually use.
Coding
Break a monolithic function into single-responsibility pieces
A 300-line function that works is still a 300-line function someone has to understand all at once to change safely.
Coding
Choose between two design directions without just picking your favorite
Liking one better and solving the actual problem are different questions.
Design
Decide what not to build this quarter, not just what to prioritize
A roadmap is really a cutting exercise wearing a prioritization costume.
Product
Turn a vague feature request into a spec engineering can actually build
Find the ambiguity before an engineer finds it for you, mid-sprint.
Product
Design a rollback plan before you ship a risky migration
A migration plan without a rollback plan is a bet you can't take back.
Coding
Get a real second opinion on an architecture decision before you commit to it
Pressure-test the choice while it's still cheap to change, not after it's load-bearing.
Coding
Decide if a feature request is really about something else entirely
The requested feature and the actual underlying need are often different things.
Product
Decide if a bug is worth a real fix or a quick patch
Patching the symptom and fixing the root cause are both valid — the wrong choice for a given bug is the expensive mistake.
Coding
Name the actual design system inconsistencies in a UI before they multiply
Catch drift while it's three screens, not the whole product.
Design
Decide your actual pricing before you guess at a number
Most early pricing is a guess anchored on a competitor's number, not a decision grounded in your actual value or costs.
Startups
Convert code from one language to another
A literal line-by-line translation between languages often compiles fine and still isn't idiomatic in the new one.
Coding
Decide if a metric improvement actually came from the feature you shipped
A metric moving up right after a launch is a correlation you want to be a story about your feature — check before telling it that way.
Product
Run a lightweight product experiment from hypothesis to decision
WorkflowMost "experiments" never had a real hypothesis, which is why they never produce a real decision.
Product
Cut a feature list down to a real MVP
A feature list only gets smaller when someone's willing to be ruthless about it — left alone, it only grows.
Product
Get a code review that catches what linters and tests miss
Style is checkable by machine. The expensive bugs live in the assumptions nobody wrote down.
Coding
Figure out why a design feels off even though it follows every rule
Checking every rule and still sensing something's wrong means the problem isn't a rule at all.
Design
Write an investor update that doesn't bury the bad news
The update that builds trust is the one where the hard part is easy to find.
Startups
Turn churn interviews into a prioritized retention plan, not a list of complaints
Ten reasons customers gave you aren't ten problems to fix in parallel.
Product
Figure out if you're solving a vitamin problem or a painkiller problem
"Nice to have" and "can't function without it" both generate polite interest — only one generates paying customers reliably.
Startups
Figure out if a design pattern is actually solving the right problem
A modal, a wizard, a dashboard — each is the right answer to some problem, and it's easy to reach for a familiar one instead of the one this problem actually needs.
Design
Take a feature from problem to a design ready for engineering
WorkflowMost design work stalls between "I have an idea" and "here's what to build," not at the sketching.
Design
Explain and rewrite a regex pattern
A regex someone else wrote (or you wrote six months ago) is often faster to replace than to actually read.
Coding
Validate a startup idea before you build anything
WorkflowEnthusiasm for an idea and evidence it's worth building are different things.
Startups
Write a product spec stakeholders actually read instead of skimming
A 12-page spec nobody finishes protects nobody when something ships wrong.
Product
Generate realistic mock test data
"Test User 1, test@test.com" doesn't catch the bugs that show up with a real name that has an apostrophe in it.
Coding
Get grilled by a skeptical angel investor
A pitch that sounds confident to a friend and one that survives a genuinely skeptical investor are tested very differently.
Startups
Run a pivot decision when the current approach isn't working
WorkflowPersistence and stubbornness look identical from the inside — the real evidence is what tells them apart.
Startups
List the edge cases that could break this validation logic
Validation logic that handles the obvious cases well usually fails on the ones nobody thought to try.
Coding
Turn scattered user feedback into a prioritized roadmap
WorkflowA pile of feedback and a prioritized set of things worth building are separated by real synthesis, not just sorting.
Product
Refactor a legacy function to modern patterns
Old code that still works isn't necessarily code worth leaving alone — it's code that's gotten harder to touch safely with each year it's avoided.
Coding
Run a design system audit and consolidation
WorkflowA design system that's grown organically for years usually has three ways to do the same thing and no one's sure which is current.
Design
Write a test plan for a change to code that has no tests yet
"Just add tests" isn't a plan when the code being changed has none to build on.
Coding
Sunset a feature without breaking the users who depend on it
WorkflowDeciding a feature should go away is the easy part — finding out who actually depends on it is where sunsets go wrong.
Product
Rubber-duck debug a problem that has no clear error message
Explaining the problem out loud, step by step, surfaces the wrong assumption faster than staring at the code again.
Coding
Run a systematic debugging session from symptom to root cause
WorkflowGuessing at fixes based on the symptom usually costs more time than isolating the actual cause first.
Coding
Scan this code for common security vulnerabilities
Code that works correctly for a legitimate user can still be wide open to one who isn't.
Coding
Write a comprehensive unit test suite for a function
A test suite that only checks the happy path proves the function works when nothing goes wrong, which was never really in question.
Coding
Write a high-impact investor intro email
A VC's inbox rewards whoever gets to the actual traction fastest — everything before that is filler they're skimming past.
Startups