AI agents for autonomous hiring workflows and talent intelligence
MakiPeople builds autonomous AI agents embedded in HR workflows, not standalone tools. The stack—TypeScript + Python + PostgreSQL on GCP, with GPT-4, BERT, DeepSeek, and PyTorch for model work—reflects a company shipping production AI at scale. Projects around psychometric calibration, bias/fairness analysis, and LLM scoring evaluation reveal a focus on making hiring decisions measurable and defensible, not just faster. Support-heavy hiring (7 of 17 roles) alongside research (3 roles) suggests Maki is scaling customer success and model robustness in parallel, a pattern typical of early-stage AI products hitting adoption friction.
Maki is an AI-powered HR platform founded in 2021 that automates talent workflows by embedding autonomous agents into recruiting and talent management. The product targets mid-market and enterprise HR teams dealing with fragmented screening, scheduling, and coordination tools. Maki captures unstructured talent signals—CVs, interview data, performance outcomes—and compounds them into predictive models for hiring, retention, and mobility decisions. The company operates across the United States, France, and the United Kingdom, with 51–200 employees based in New York City.
MakiPeople uses GPT-4, BERT, and DeepSeek for language understanding, PyTorch for model training, and Hugging Face for pre-trained transformers. They also employ ClickUp, n8n, and Apache Airflow to orchestrate agentic workflows.
Active projects include predictive retention analytics, bias and fairness in AI scoring, psychometric model calibration, LLM evaluation, semantic embedding benchmarking, and security tooling (CI/CD, IaC). They also focus on voice AI security and automation for care workflows.
MakiPeople's technology stack, projects, and hiring signals are inferred from public hiring and company data — career pages, public listings, and company web presence — then clustered and de-duplicated. Figures are estimates that refresh over time. Read our full methodology →
This is not an official vendor or customer list. It is a technology-adoption signal inferred from public data, intended for B2B research.