What a Great AI Engineer's Portfolio Actually Looks Like

Forget GitHub stars. Here's what to actually look for when a candidate shows you their past AI work, and the questions that separate builders from tutorial-followers.

Elena Voss·Head of AI Delivery, Aiporate··7 min read·Share on XLinkedIn

Key takeaways

  • Weight evidence of measurement, evals, before/after comparisons, cost tracking, far above polish or star counts.
  • Look for at least one project that ran in production long enough to have a failure story, not just a demo that worked once.
  • A strong portfolio shows tradeoff reasoning: why this retrieval approach, why this model size, why not the more obvious choice.
  • Ask to see the worst output the system ever produced; candidates who can show you this have actually operated something real.
  • Depth on one real project beats breadth across ten tutorial clones, every time.

Star counts, demo polish and a list of impressive-sounding model names tell you almost nothing about whether someone can build an AI system that survives contact with real users. What tells you something is far less glamorous: did they measure quality before claiming it improved, did they make a deliberate tradeoff and explain why, did anything they built actually run in production long enough to break. A great portfolio reads more like an incident log than a highlight reel.

The signals that don't actually tell you much

It's easy to over-index on things that are visible and easy to compare, and that's exactly why they're weak signals: they're the easiest things for a candidate to optimize for appearance rather than substance.

  • GitHub star counts, they correlate with marketing and timing more than engineering quality.
  • A long list of model and framework names, name-dropping is free and doesn't demonstrate judgment.
  • Polished demo videos, a demo is optimized to work once, on curated input, which is precisely what production never gives you.
  • Number of projects, ten shallow tutorial clones say less than one project taken to production and maintained.

What to look for instead

  1. 1Evidence of measurement: did they define what 'good' meant before or after they built the thing? Look for an eval set, a metric, a before/after number, anything that shows quality was managed, not assumed.
  2. 2A production lifespan: has anything they built run long enough, with real users or real data, to have needed a fix after launch? A project that's still running six months later teaches you more about the builder than one that shipped and was never touched again.
  3. 3Tradeoff reasoning: can they explain why they chose one retrieval strategy, model size or architecture over an equally plausible alternative, and what it cost them? Judgment shows up in the roads not taken.
  4. 4A failure they can describe in detail: what broke, how they found out, what they changed. This is the single highest-signal thing you can ask for.
  5. 5Ownership of the unglamorous parts: logging, cost monitoring, a rollback plan. These rarely make it into a portfolio deck, which is exactly why asking about them directly is so effective.

How to actually run the portfolio review

Don't just read it, interrogate it. Pick the one project the candidate seems proudest of and spend fifteen minutes going deep rather than skimming five projects shallowly. The goal is to find the edge of their real understanding, which is usually one or two follow-up questions past whatever they volunteered first.

  • "Show me the worst output this system ever produced, and what you did about it."
  • "What would you build differently if you started this project again today?"
  • "Who else used this, and how do you know whether it actually helped them?"
  • "What did you decide not to build, and why?"

Weak portfolio vs strong portfolio, side by side

DimensionWeak portfolioStrong portfolio
Evidence of qualityClaims it 'worked well'Cites a specific eval score or before/after comparison
Production exposureDemo, ran once, never touched againRan for real users, was maintained and fixed over time
Tradeoff awarenessOne approach presented as obviously correctExplains what else was considered and why it lost
Failure handlingNo failures mentionedA specific failure, root cause and fix
Cost/latency awarenessNever mentionedDiscussed as a real constraint that shaped decisions
What the difference actually looks like

Frequently asked questions

What if a candidate's best work is confidential and they can't share code?

That's normal and fine, ask them to walk through it verbally with the same depth you'd want from a code review: what they measured, what broke, what they'd change. Depth of explanation matters more than being able to see the repo.

How much should side projects and hackathon work count?

They're useful for gauging curiosity and initiative but shouldn't be weighted like production experience. A hackathon project that never had real users can't have a failure story, and the failure story is where most of the signal lives.

Is it fair to judge junior candidates by production lifespan?

Adjust the bar, not the criteria. A junior candidate might not have owned a production system yet, but you can still look for measurement instinct and tradeoff reasoning in whatever they have built, a strong junior often shows the same instincts at smaller scale.

Head of AI Delivery, Aiporate

Elena has spent 12 years building and embedding AI and data teams inside B2B SaaS companies, from first pilot to enterprise-wide platform. At Aiporate she leads how forward-deployed talent is matched, onboarded and shipped to production.

Need the team to make this real?

Describe your need in plain English, get the exact hire, forward-deployed talent or a fractional leader, vetted and matched in 72 hours.

Scope your need →

Keep reading

The Weekly Brief

Intelligence for building AI-native organizations.

One email a week: the sharpest thinking on AI hiring, infrastructure, teams and strategy, for the people building the future of work.

Join operators, founders and CTOs. No spam, unsubscribe anytime.