We staff projects, not positions.
Alot Labs runs on a small permanent core, per-project build pods, and a bench of Claude-certified engineers. Every open seat below belongs to a named lab project. When it ships, the pod wraps and the credit is yours.
The Core
The permanent center: the founder plus the firm’s standing agent infrastructure — shared memory, build standards, CI, evaluation harnesses. The Core owns client relationships, final review, and everything career-shaped: credit, feedback, the next engagement. Pods dissolve; the Core remembers.
Build pods
The delivery unit, and the hiring unit. One pod per lab project: a pod lead, sometimes a second builder, and a named crew of Claude agents carrying the execution volume. A pod owns one outcome, states its human-agent ratio up front, and wraps when it ships. Leading one is a steering role, not a management rank.
The Bench
The roster pods are staffed from: Claude-certified engineers who have crewed a pod or passed a paid trial sprint. Staffing goes bench-first, open applications second. Your record here is a credit list — certifications plus pods shipped — and it carries to the next charter.
A typical pod runs one or two humans with four to eight agents. Humans own judgment — architecture, acceptance, the call on what ships. Agents own volume — drafts, tests, migrations, docs. The agents are listed on the pod charter because they do real work; the humans are listed because someone has to answer for it.
One seat per project. Read the project on the lab page before you write to us.
STG
Agent-driven QA. Collapse ticket, doc, and test into one executable object and make it hold up in CI. You direct the agent crew that operates it.
nearby
Outreach infrastructure past phase zero: territory scanning, diagnosis pipelines, and the staged-autonomy rules that keep it honest.
flywheel
A deterministic finance core with an AI advisor on top. Tested math, versioned tax config, local-first — and three named limits to close out.
holymolypoly
Market-data plumbing and risk controls: live feeds, sizing rules, kill-switch reliability, and the paper-trade evidence that decides if it ever goes live.
jaanal
Simulation and evaluation work: agent reasoning loops, metrics, and the evidence that earns the fine-tune.
LPP
A company as software: agent roles, an operating cycle, a constitution enforced in code. The design is done; you build Phase 2.
academy
Stand up the platform behind the academy–talent–services flywheel. The mastery-gate curriculum engine ships first.
RAW
An AI-native operating stack for a real-world venture: skills, decision logs, and the ops discipline that keeps a company honest.
banks
Agent-economy plumbing: harden the ledger, extend escrow and credit lines, take the hash-chain audit further.
compliance
From spec to on-prem agent: camera discovery, a registry with purpose and retention, and an audit trail a regulator accepts.
A Claude certification from Anthropic — every human seat, no exceptions. Something you shipped that we can read. And proof you can direct agents rather than babysit them: show us a build where the agents did the volume and you made the calls.
Pods are contract engagements with a chartered end date. Ship one and you are on the Bench for the next.
The button opens your mail app. If nothing happens, use the form below, or write to careers@alotlabs.com from any client — pod name in the subject line, certification and your GitHub or LinkedIn in the body.
Apply from right here.
Name the pod, prove the certification, drop your GitHub or LinkedIn. Same inbox as the button, same reply.