This week’s deeper signal: regulated, document-heavy industries are growing headcount while rolling out GenAI. That combination is pulling hiring toward engineers who can ship fast and build for security, audit, and reliability.
The market, by the numbers (new data points)
UK hiring is cautious: job postings are 19% below pre-pandemic levels, unemployment hit 5% (Q3 2025), and job-to-job moves were reported at 708,000, down almost 30% from the late-2021 peak.
Legal is still adding jobs: the US legal sector added 5,500 jobs in Jan 2026, the 18th consecutive month of gains, reaching about 1.24 million total jobs, up 8.8% versus five years ago.
GenAI adoption inside UK law is already mainstream: 61% of lawyers say they use GenAI at work, up from 46% in Jan 2025. Those with no plans to adopt fell from 15% to 6%.
But implementation is uneven: only 17% say AI is fully embedded in strategy and operations, while 47% think AI will transform billing models.
Retention signal: 18% of private practice lawyers and 19% of in-house counsel would consider leaving if their org doesn’t invest in AI, rising to 26% at large firms.
Translation: hiring is tighter in general, but high-stakes workflows are still investing, and they care about “boring” engineering excellence: access control, observability, incident response, and accuracy under audit.
Deeper this week: a 40-minute screen for “regulated AI” engineers
If you want to evaluate Full Stack or Python engineers for AI features in sensitive workflows, test three measurable instincts:
1) Security and access model (10 minutes)
Give a short scenario: “Users upload confidential documents. Some matters are restricted to a subset of users.”
Ask for:
a simple authorization model (matter-level ACLs, roles)
one failure mode they’d prevent (leaky search results, cached responses, shared embeddings)
Score: can they articulate data boundaries clearly, not just “we encrypt stuff.”
2) Ingestion and retrieval performance (15 minutes)
Provide a basic constraint: “A matter has 50,000 documents. You need search + summarisation.”
Ask:
their ingestion pipeline stages (dedupe, OCR if needed, chunking, metadata, embeddings)
how they’d keep it fast (batching, async workers, backpressure, retries)
what they’d log (ingestion latency, failure rate, queue depth, retry counts)
Score: do they think in throughput, failure rates, and observability.
3) Quality and auditability (15 minutes)
Ask for an evaluation plan:
“What does ‘good’ look like in week 1?”
“How do you detect silent failures?”
Score on whether they propose:
test sets (even small, curated)
traceability (citations to source docs, per-answer evidence)
monitoring metrics (answer rate, fallback rate, confidence proxies, user feedback loops)
This screen works because it forces candidates to demonstrate how they’d build AI features that survive scrutiny, not just demos.
Quick Python watch
Only the stuff that can realistically affect production this week:
Django security releases: Django issued 6.0.2, 5.2.11, and 4.2.28 as security updates. If you run Django in production, this is a patch window, not a “someday.”
Ruff: Ruff 0.15.x landed in February 2026, with formatter changes and a migration guide. If your CI or formatting is pinned, check before you auto-bump.
Python releases: Python source releases show 3.13.12 and 3.14.3 published on Feb 3, 2026. Worth tracking if you maintain base images or buildpacks.
Job of the week
Founding Full Stack Engineer, YC-backed, AI for litigators (London)
YC X25 London startup building an AI associate for high-stakes litigation. They’re already working with top UK and US law firms on workflows like drafting submissions, analysing judgments/orders, summarising transcripts, and locating key evidence fast.
Why it’s exciting for engineers
Real end-to-end ownership: product, infra, prompting, UX, and performance
Hard constraints that make engineering meaningful: security, reliability, speed, and sensitive data
Agentic workflows on messy real-world documents, not toy demos
Stack
Python backend, Next.js + TypeScript frontend, PostgreSQL
Comp
£100k–£130k base
0.50%–1.00% equity
London, minimum 3 days/week
If you want the full brief and intro, reply and I’ll share details privately.
Hiring? Contact
Josh Smith
LinkedIn: https://www.linkedin.com/in/python-recruitment/
Email: [email protected]
Phone: 01727 225 552