Aquin (Aquin Labs) is an AI research organisation working on interpretability and applied prediction. Its core project, Aleph, is positioned as a universal predictor for applied ML, data science and annotation work, producing schema-bound outputs such as probabilities, labels, boxes and scores that programs can consume. Alongside it the lab publishes research on attention circuits, weight editing and self-learning developer agents.
Workable is an agentic recruiting and HR platform. A native AI agent sources passive candidates from 400M+ profiles, screens every applicant against your job criteria and builds vetted shortlists inside a complete ATS, while onboarding, performance, time off, employee records and reporting run on the same system of record.
Aquin (Aquin Labs) is an AI research organisation working on interpretability and applied prediction. Its core project, Aleph, is positioned as a universal predictor for applied ML, data science and annotation work, producing schema-bound outputs such as probabilities, labels, boxes and scores that programs can consume. Alongside it the lab publishes research on attention circuits, weight editing and self-learning developer agents.
Universal predictor (Aleph)
A single base that handles tables, media, annotation and decision tasks, with the option to fine-tune toward your specific data.
Schema-bound outputs
Outputs are typed and scored rather than free text, so downstream programs can act on results and escalate low-confidence cases.
Interpretability research
Treats attention as a routing map from a single forward pass, isolating a small subgraph aimed at the answer and testing it under ablation.
Weight editing experiments
Agentic weight-editing experiments that localise causal layers, apply rank-one MLP updates and validate then roll back steps when checks fail.
優勢與優點
+Research-grade interpretability work published openly alongside the applied product
+Outputs designed for programs rather than chat, which suits automation pipelines
+Explicit confidence scores make it practical to escalate uncertain predictions
+One base is claimed to cover tables, media, annotation and decisions
限制
–Aquin Labs is a research organisation, so most output is papers and interfaces rather than a polished consumer app
–Aleph is described as a predictor for programs and teams, not a general chat assistant
–No transparent public price list was found for commercial engagements
–Material is heavily technical and assumes familiarity with ML evaluation work
Workable
Workable is an agentic recruiting and HR platform. A native AI agent sources passive candidates from 400M+ profiles, screens every applicant against your job criteria and builds vetted shortlists inside a complete ATS, while onboarding, performance, time off, employee records and reporting run on the same system of record.
AI candidate sourcing
The agent continuously searches 400M+ profiles, identifies strong fits before they apply and sends each a personalised outreach message for every open role.
Automated screening and engagement
Every applicant is scored against up to 14 job criteria with visible reasoning, then messaged directly to confirm interest and close information gaps before reaching a recruiter.
Complete ATS underneath
Requisitions, pipelines, interview kits and scorecards, interview scheduling, offer letters with e-signatures and integrated background checks in one platform.
Job ad syndication and careers pages
Publish jobs to 200+ boards in 100+ countries and build branded career sites with a drag-and-drop editor, plus LinkedIn and Indeed Apply integration.
優勢與優點
+Native AI agent does real sourcing, screening and candidate messaging, not just keyword ranking
+Runs on top of a complete ATS rather than being a bolt-on tool with limited context
+
Hiring context carries through to onboarding and HR, so talent data compounds over time
+Open and MCP-native: bring Claude, ChatGPT or your own agents while Workable stays the system of record
+Strong distribution via 200+ job boards, LinkedIn and Indeed preferred partnerships
+Trusted by 35,000+ teams, with 270+ integrations and mobile-first candidate experience
限制
–Pricing scales with headcount and module selection, so costs grow as you add HR features
–Full value depends on adopting both the ATS and HR modules, which is a bigger rollout for small teams
–Very feature-rich, so administrators need time to configure workflows, permissions and approvals
–Partner job-board and integration advantages vary by region and plan tier
#1
