Employment Screening

You're allowed to verify a candidate's background. You're not allowed to do it however you want. Background checks sit at the intersection of three regulations that rarely agree with each other: data privacy, pay transparency, and local labor law. Get the intersection wrong, and the check itself becomes the liability. Here's what's actually possible, country by country — and what gets companies fined. Germany You can't run your own criminal record search. Only the candidate can request their own Führungszeugnis (police clearance certificate), and you can only require one when it's directly tied to the role — finance, childcare, critical infrastructure. Make it a blanket policy for every hire, and you're outside GDPR's proportionality principle. United Kingdom Criminal record data is "special category" under UK GDPR. Consent alone doesn't make it lawful — the power imbalance in an employment relationship means you need a separate legal basis under Schedule 1 of the Data Protection Act. References aren't even a legal requirement outside regulated sectors like education and care. Brazil LGPD requires specific, written, informed consent before any check — no blanket authorizations. Credit history is off-limits unless the role carries real financial responsibility. Criminal record checks are restricted to roles where security is legally mandated. India The DPDP Act requires consent that's free, specific, and revocable — not a clause buried in an offer letter. Data collection has to be proportional to the role, full stop. Enforcement tightens further in 2027, so the standard to build to now is already set. United States FCRA governs the mechanics of the check itself: separate written disclosure, written authorization, and adverse action notices with dispute rights before you reject anyone based on a report. Layer on 37+ states with ban-the-box laws delaying criminal history questions until after a conditional offer, and 18+ states now banning salary history questions outright. EU-wide The Pay Transparency Directive's transposition deadline passed in June 2026. The direction is set: asking candidates about pay history is headed toward a bloc-wide ban, and pay ranges will need to be disclosed upfront in the hiring process. The pattern across all five: consent isn't a formality, "relevant to the role" is doing all the legal work, and pay history is becoming untouchable almost everywhere you hire. Best practice isn't "check everything." It's checking the right thing, in the right country, with the right paper trail — before you extend the offer, not after a regulator asks. This is exactly the kind of call ONE, our compliance agentic companion, is built to help you make — country by country, hire by hire. Laws here move fast and vary by state, sector, and role. Treat this as a starting point for your own legal review, not a compliance sign-off.

I’ve reviewed thousands of job applications from academic scientists looking to move into biotech startups. Here’s how the best applications stood out ⤵️ Sharing this for folks graduating from PhDs this year or thinking about a change - it’s still a tough market out there, but one that’s hopefully improving! _______ 1️⃣ Show how your personal values align to the company mission. Why? Startups want to change the future. Demonstrate you’ve been independently working towards that same future → this indicates you’ll work hard & find the day to day meaningful. How? Example, for a company developing phages to treat antibiotic resistant bacteria: ✅ My PhD research focused on optimising a gene therapy for children suffering from grey platelet syndrome. During that time, I volunteered in the pediatrics ward. I am motivated by improving health outcomes for the most vulnerable. ❌ Having finished my PhD, I am looking to make the jump into industry. _______ 2️⃣ Directly explain how your scientific expertise can solve the startup’s problems. Why? This shows your ability to connect the dots between “the company problem that needs to be solved” and “the impact I can have.” Startup MVPs have proactivity in spades. How? Example, for a company developing cultured meat: ✅ A big problem for the cultured meat industry is developing immortalised, scalable cell lines. As a genetic engineer, I can generate cell lines capable of feeding millions of people. ❌ My 6 years of experience with mammalian cell culture and background in genetic editing make me a great fit for your company. _______ 3️⃣ Incorporate metrics (beyond publications!) into your resume. Why? Publications = academic currency. Scientific breakthroughs allowing a company to get profitable and survive = startup currency. Publications require detailed science capable of getting past peer-review. Startups require time-boxed, outcomes-oriented science. That’s really different! Metrics indicate you already understand that shift in mindset - and no matter what your project focused on, you can frame it in terms of startup-relevant metrics. How? ✅ Supported two summer students to achieve xyz outcome in three months ✅ Generated 5 novel immune complexes in 2 months ✅ Achieved XYZ while dropping experiment costs by 20% ❌ Conducted a research project analysing how XYZ ❌ Published in a prestigious journal. _______ 4️⃣ Show - don’t state - your communication & collaboration skills. Why? These skills are 10x more important when working at a fast pace with people from different professional backgrounds. How? ✅ Three-minute thesis contest ✅ Industry/startup work experience ✅ Engagement with an entrepreneurship community ✅ Cross-discipline collaboration ✅ A well-written career summary connecting the dots between your skills & the value you can bring to the company. As always, builds or add-ons welcome: I made some of these mistakes when I first graduated from my PhD, you don’t have to 😉

📢 UNEP FI just updated their fully open-source Sector Impact Matrix, identifying sector-specific risks and opportunities linked to sustainability issues across 1,000+ sectors! The Matrix was developed with the Impact Management Platform (IMP) and peer reviewed by leading standard setters, banks, investors, and asset managers.  This is helpful because most materiality assessments treat sectors as one size fits all. A tool like this one that maps impact drivers and affected parties for a variety of sectors saves valuable time.  𝗪𝗵𝗮𝘁 𝘁𝗵𝗲 𝗺𝗮𝘁𝗿𝗶𝘅 𝗶𝗻𝗰𝗹𝘂𝗱𝗲𝘀:   • Positive and negative impacts on people, society and the natural environment mapped for 1,000+ sectors    • All sector-impact associations fully explained and referenced   • New data points included for each sector-impact association: affected parties, impact drivers, value chain position, associated risks and opportunities   • Powerful new interactive search functionalities  The Matrix aligns with numerous sustainability frameworks including TNFD, TISFD, ESRS, GRI, and ENCORE. See the full list below.  𝗛𝗲𝗿𝗲’𝘀 𝘄𝗵𝗮𝘁 𝘆𝗼𝘂 𝗰𝗮𝗻 𝗱𝗼 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗠𝗮𝘁𝗿𝗶𝘅:   • Financial institutions and investors can strengthen materiality assessments, disclosures, and client or investee screening and due diligence processes    • Corporates can improve product and service development    • Data providers can supplement sector analysis, benchmarking and ratings    • Sustainability professionals can use this for internal research and to support efforts towards interoperability  Sustainability issues are complex and interconnected, the Matrix is an essential resource in navigating these.  Check it out here: https://lnkd.in/eDRkAHys United Nations Environment Programme Finance Initiative (UNEP FI) Impact Management Platform #sustainablefinance #decisionmaking #materialityassessment #impactmanagement  

I rejected a perfect candidate last year. Not me personally. My AI screening tool did. 𝐈 𝐝𝐢𝐝𝐧𝐭 𝐞𝐯𝐞𝐧 𝐤𝐧𝐨𝐰. 3 first-author papers on reinforcement learning. 200+ Google Scholar citations. Stanford-funded research. The kind of profile recruiters dream about. The AI scored them 34 out of 100. Why? Their CV said "statistical learning systems" instead of "machine learning." Thats it. One synonym. The tool couldnt make the connection. I only found out because I manually reviewed the reject pile on a hunch. 47 profiles deep into an 8-hour sourcing session. If I hadnt looked, my competitor would have placed them. (Most recruiters dont know their AI screening tools cant distinguish between technical synonyms — and theyre making decisions on hundreds of thousands of applications.) This isnt a one-off. Across 28 businesses, Ive documented the same pattern: AI systematically rejects candidates with non-linear careers, unconventional project descriptions, or terminology that doesnt match the job spec word-for-word. 19% of organisations using AI in hiring admit their tools screen out qualified people. SHRM published that number. The real number is higher. Most teams dont check. Heres what I changed: every AI-screened shortlist gets a human verification pass. Every one. I built a prompt engineering framework for JD analysis so the AI actually understands context before it scores. Time-to-screen dropped 60%. Not because the AI got better. Because a human catches what it misses. The EU AI Act classifies every CV screening tool as high-risk. August 2026. 115 days. Fines up to 35M euros. Most recruiting teams still cant explain what their AI tools actually do. Do you manually check your AI-screened shortlists, or do you trust the scores? Save this before your next screening audit.

✅ All Properties Pass. Formal Signoff Done? Not so fast. Here's a story every formal engineer should relate to. You've been working on a formal testbench for weeks. All assertions are green. No counterexamples. Everything is passing. You think: “We nailed it. Time to declare formal signoff.” But wait—what if your verification environment itself is flawed? What if the properties are incorrect or incomplete That’s the trap. One that many experienced engineers have fallen into. 🔁 Redundant Properties to the Rescue Imagine this: You wrote an assertion that every req must be followed by an ack within 3 cycles. It passes. To double-check, you also model a small helper code that tracks the req and raises an error if ack doesn’t arrive in time. This helper code catches an issue. Turns out your original property had a subtle bug. It passed, but for the wrong reason. 📦 The Classic FIFO Case Your FIFO is verified. Data in, data out, all looks good. But when you write a second property—using a scoreboard to track order—you notice discrepancies. Even worse, your cover properties never trigger. Why? Because the stimulus never exercised a full-to-empty scenario. So your assertions were passing vacuously—never truly verifying anything. 🟩 Cover Properties Covers help ensure that you are hitting important scenarios in design. For example: Cover a scenario where fifo_full is high, but wr_en is still asserted. Cover a full-to-empty transition to verify realistic usage. If a cover doesn't hit, something may be wrong in the environment—or worse, the assertion may be checking nothing at all. Using redundancy and coverage in property verification dramatically reduces the risk of missed bugs and builds trust in your formal signoff. 💬 What techniques do you use to ensure your properties are correct? Let’s share and learn from each other! #FormalVerification #VLSI #SystemVerilog #EDA #DesignVerification #RTL #FIFO #Assertions #StaticSignoff #Semiconductors

Background checks. Sensitive data. Zero DPDP compliance. The most sensitive personal data comes from your hiring process. 📌 Criminal records. 📌 Financial history. 📌 Past employment. 📌 Address verification. 📌 Education certificates. And almost no Indian company has a DPDP-compliant process for any of it. Here is the legal reality your HR team doesn't know: Your company = Data Fiduciary. Your BGV vendor = Data Processor. Your candidate = Data Principal with enforceable rights under DPDP. Every obligation that applies to your customer data — applies here too. The 5 gaps I find in almost every BGV process I review: 1️⃣ Consent was never properly obtained. Most companies collect a generic clause inside the offer letter. Under DPDP — consent for a background check must be specific to that purpose, informed about what will be verified and with which sources, and separate from the employment acceptance. "I accept this offer" is not consent to a criminal record check. 2️⃣ No signed DPA with the BGV vendor. You have a commercial agreement with your BGV vendor. Under DPDP — that vendor relationship requires a Data Processing Agreement with breach notification timelines, deletion obligations, sub-processor controls, and Data Principal rights flowing down. A commercial agreement and a DPA are not the same document. 3️⃣ Candidate rights are completely unaddressed. Under DPDP, your candidate has the right to access what data was collected about them, from which sources, and what the report concluded. Most HR teams have no process for this. No one has asked before — but it is now a legal right, not a courtesy. 4️⃣ BGV reports are retained indefinitely. The candidate joined — or didn't. The report is still in your HRMS, your email, your recruiter's drive — years later. Under DPDP — personal data must be deleted once the purpose is fulfilled. The purpose of a background check is the hiring decision. Once made — the legal basis for retaining the report ends. 5️⃣ Cross-border transfers nobody mapped. Most BGV vendors verify employment and academic records through international databases. That is a cross-border data transfer. Under DPDP Section 16 — your company is responsible for it. Not your vendor. Does your BGV vendor's contract specify which countries your candidate's data flows to? _____________________________ The background verification industry processes thousands of sensitive personal data records every month in India. Almost none of it is DPDP-compliant. And the liability doesn't sit with the BGV vendor. It sits with the company that initiated the check and is the Data Fiduciary. Does your company have a signed DPA with your BGV vendor? ___________________ I help companies build DPDP-compliant hiring data processes — from candidate consent to vendor DPAs to rights response frameworks. Book 1:1 call to find out where you stand. (Link in comment.)

ICLR'26 has decided to 𝗱𝗲𝘀𝗸-𝗿𝗲𝗷𝗲𝗰𝘁 papers with 𝗵𝗮𝗹𝗹𝘂𝗰𝗶𝗻𝗮𝘁𝗲𝗱 𝗿𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲𝘀 generated by LLMs. That raises a practical question for every author: How do we verify citations reliably? We’re excited to share our new paper, 𝗕𝗶𝗯𝗔𝗴𝗲𝗻𝘁, an agentic citation verification framework designed to make reference checking auditable BibAgent traces where a claim is supported, surfaces evidence spans, and reports confidence rather than guessing. When a cited paper is behind a paywall, it can switch to a community-based “evidence committee” approach: collect downstream open-access citers, distill what they attribute to the paywalled work, and decide with consensus—or abstain if evidence is insufficient. We also propose a unified miscitation error-code taxonomy and release 𝗠𝗜𝗦𝗖𝗜𝗧𝗘𝗕𝗘𝗡𝗖𝗛, a large cross-disciplinary benchmark of miscitation cases. If you’re building LLM writing assistants, submission pipelines, or research integrity tooling, this is meant to be a step toward: draft fast → verify rigorously → publish faithfully. Ultimately, we hope this research helps 𝗳𝗮𝗰𝗶𝗹𝗶𝘁𝗮𝘁𝗲 𝗳𝗮𝗶𝘁𝗵𝗳𝘂𝗹 𝗮𝗻𝗱 𝘁𝗿𝘂𝘀𝘁𝘄𝗼𝗿𝘁𝗵𝘆 𝘀𝗰𝗶𝗲𝗻𝘁𝗶𝗳𝗶𝗰 𝗽𝘂𝗯𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀—so that emerging agents for scientific discovery can build on literature that’s genuinely grounded, not citation-shaped. Paper link: arxiv.org/abs/2601.16993 #AI #LLMs #ResearchIntegrity #OpenScience #NLP #ScientificDiscovery #TrustworthyAI

Clearly, the approach to job application assessment will have to drastically change as unmanageable waves of applications - a lot of them being AI generated - are now hitting recruiters. I'm seeing fairly common job openings now often gathering over 1,000 applications within a day of being posted online. A lot of them submitted at odd times, usually the second when the job posting is scrapped by an AI. With the emergence of AI in recruitment, many candidates have put their job search in the hands of an AI agent. These AI bots will scrape relevant job postings, analyze requirements, and then generate an entire application, with cover letter, CV, etc. It can apply to a few hundred roles daily, and will continue doing so until it's switched off, which could never happen as candidates may want to continuously test the market. When looking at these applications, cover letters look exactly the same, using the now familiar AI generated phraseology, and CVs are so similar that they are often assumed by recruiters to be originating from scammers and other impersonators. From a volume perspective, it's impossible for a human to effectively find the right skills and candidates from a stack of 1,000 applications. And old school keyword based assessment are no longer effective because AI bots are peppering generated CVs with keywords founds in the job posting, when they are not copying and pasting entire sections of the job spec. What are the solutions? In the short term, from my perspective, beyond using the usual behavioral signal screening (e.g., time spent on job descriptions, etc.) and adding friction to the application process (e.g., limit how many roles someone can apply to, ask simple thoughtful questions to make sure they’ve put in real effort), I believe firms should start rethinking assessment relative to skills and roles. More specifically, 1). Embed more pre-qualification assessments or simulations before a recruiter spends time reviewing the application (e.g., role-relevant assessments, skills quizzes, situational judgment tests, etc.) early in the funnel to make sure that candidates meet basic requirements. 2). Build talent pools (vs focusing on requisition based recruitment). Basically, proactively building groups of interested and qualified people we can reach out to when the time is right. This requires more planning and a mature TA org though. 3). Ultimately, firms should actively embrace resume-free screening for selected roles. Basically, we skip the CV entirely and ask for proof of ability (e.g., a test) etc. In the long run though, I'd be interested in seeing how the most advanced assessment suppliers will innovate in the areas of candidate identity, skills & reputation portability and recruiter-side AI used to contextualize fit (e.g., using career trajectory, digital footprint, etc.) #JobApplication #CandidateAssessment #TalentAcquisition #Skills https://lnkd.in/eMGpHD2c