Why this matters
Most bad base44 hires were predicted by visible red flags the hiring manager noticed and ignored. The flags are not subtle. The pattern is consistent: a candidate misrepresents expertise on first contact, the hiring manager wants the engagement to work, the candidate clears a low-evidence interview, and the engagement fails at week two. The fix is administrative — write the flags down, walk through them, and walk away when two or more appear.
Who this is for
- Hiring managers under pressure to ship and tempted to skip vetting
- Founders who got burned on a previous hire and want a structured filter
- Procurement teams writing vendor evaluation criteria
- Operators screening candidates from Upwork, Fiverr, or other marketplaces where signal is noisy
- Anyone tempted to overlook one obvious flag because the hourly rate is cheap
The 12 red flags
1. No public portfolio of shipped base44 apps
The candidate cannot produce a single live URL of a base44 app they shipped. They offer screenshots, descriptions, "I worked on it but it is under NDA" — anything but a working URL.
What it predicts: the candidate has not shipped meaningful base44 work. The "NDA" answer is acceptable for one app, not all of them. See the vetting checklist item #1.
2. Refusal to sign a mutual NDA
The candidate balks at signing your standard mutual NDA, or insists you sign their template instead, or demands one-way protection without reciprocal obligation.
What it predicts: discovery-related contract disputes downstream. Mutual NDAs are standard at every level of the industry. Any version of "we don't really do NDAs" is a walk-away.
3. Vague pricing or "we'll figure it out"
The candidate cannot quote a price for a defined scope. They want to start work and "see where it goes." They quote a range like $5K-$50K with no narrowing criteria.
What it predicts: budget overruns. Senior engineers price scoped work in 30 minutes. Candidates who cannot are either inexperienced (priced wrong) or playing for hourly creep.
4. No platform-specific knowledge under questioning
You ask the candidate to describe credit-burn loops, function routing 404s, the July 2025 SSO bypass, AI-agent regressions, or any documented base44 failure mode. They answer with generalities ("I would check the logs"), defensive deflection ("I would research that as needed"), or admit they have not encountered it.
What it predicts: a generalist charging specialist rates. The platform-specific failures are well-documented; an engineer who has shipped 50+ base44 apps has seen at least three of them.
5. No references or only curated friend references
The candidate cannot produce references, or produces one (always one) reference who turns out to be a friend, family member, or co-bootcamp graduate.
What it predicts: no production engagements at the seniority claimed. Senior engineers have a network of past clients willing to be referenced. The absence of that network is the signal.
6. Hourly billing with no cost cap on scoped work
You describe the scope. You ask for a quote. The candidate proposes hourly billing with no upper bound, on work a senior engineer should be able to estimate.
What it predicts: budget overruns. Hourly without a cap is the canonical mechanism for $3K engagements becoming $9K engagements. See the contract cluster page for cap structure.
7. Missing or evasive on IP assignment
The candidate's contract template has no IP assignment clause, or has an "on delivery" clause that lets them retain code if invoices are disputed, or refuses to negotiate to "on payment."
What it predicts: leverage problems if the engagement goes sideways. IP-on-payment is a non-negotiable for any production work. A candidate who pushes back on this is signaling intent to retain leverage.
8. Evasive about subcontracting
You ask whether work will be subcontracted. The candidate says "we may bring in additional resources as needed" without specifying who, where, or under what disclosure.
What it predicts: undisclosed offshore handoff. The pattern: the candidate who interviews sells the engagement, then hands the code work to a different engineer at a fraction of the rate. Code quality drops; the candidate keeps the margin.
9. Generic code samples that could be from any project
The candidate provides a code sample. It is generic — a basic CRUD endpoint, a hello-world function, a copy of a tutorial. There is nothing base44-specific about it.
What it predicts: the candidate has not written significant base44 code. Specialists provide samples that show platform-specific knowledge — function-routing patterns, credit-burn mitigation, auth flows. Generic samples reveal the gap.
10. Zero GitHub presence
The candidate has no GitHub, or has an account with one fork from three years ago and no other activity.
What it predicts: combined with another flag, lack of shipping. In 2026 senior engineers have at least passive GitHub presence. Zero presence does not always mean misrepresentation; combined with no portfolio or no references, it does.
11. Opposes written SOW
The candidate prefers verbal agreements, "lightweight" contracts, or a single-paragraph SOW that omits acceptance criteria, milestones, and termination terms.
What it predicts: scope dispute downstream. A written SOW is administrative friction designed to prevent disputes. Candidates who oppose it are either inexperienced (do not understand the value) or planning to dispute scope.
12. Demands 100% payment upfront
The candidate insists on full payment before starting work, or before any milestone is delivered.
What it predicts: cash flow problems on the candidate's side, or intent to disappear. Standard payment structures (30/30/30/10, 50/50, 20/40/40) protect both sides. 100% upfront eliminates buyer leverage. Walk away unless the engagement is sub-$1,500 with a known contractor.
How many flags is a walk-away
| Flags counted | Action |
|---|---|
| 0 | Proceed to SOW |
| 1 | Probe deeper, ask the candidate to address it explicitly |
| 2 | Walk away |
| 3+ | Walk away immediately |
The reason 2 is the threshold: any single flag can have an explanation that makes sense in context. Two flags consistently predict a bad outcome regardless of explanations. Hiring managers who relax this rule under deadline pressure regret it within the first sprint.
Some flags are walk-aways by themselves: NDA refusal (#2), missing IP assignment (#7), 100% upfront demand (#12). Each of these is structurally incompatible with a healthy engagement.
Trade-offs and pitfalls
The dominant pitfall is rationalizing a flag because the hourly rate is cheap. Cheap rate plus three flags is the most expensive engagement you will run, because the recovery cost when the engagement fails dwarfs the savings. Walk away.
The second pitfall is over-weighting good first impressions. A polished website, a professional portfolio site, and a smooth sales call do not substitute for the underlying flags. Style is uncorrelated with shipping ability.
The third pitfall is hiring against your gut. If something feels off — even if you cannot articulate which flag it is — note it and slow down. The flags are post-hoc rationalizations of the same pattern recognition.
How Base44Devs fits in
Base44Devs publishes prices, signs mutual NDAs by default, names the engineer on every SOW, assigns IP on payment, and offers money-back if a sprint cannot resolve the issue. None of the 12 flags apply. Order an audit for the lowest-risk first engagement, or book a free 15-minute call to scope a sprint.
Related options
- 32-point vetting checklist — the structured battery the flags map onto
- Hire a base44 freelancer — engagement structure for clean candidates
- Base44 specialist vs Fiverr — where these flags appear at higher density