talent-intelligence

Discover emerging AI talent before they become obvious

Premium CPU-only AI ranking system for Redrob India Runs Track 1 Intelligent Candidate Discovery

StepIn：面向零基础学生的实践优先 Learner Agent 平台，从真实简单任务开始，通过反馈、修订、迁移与证据链形成可验证能力。Practice-first learner agent for real-work simulation and capability evidence.

AI 竞对人才情报：从竞争对手公开招聘官网抓取上新岗位、叠加最新人才情报新闻，反推人才策略并沉淀为飞书多维表格仪表盘。Agent Skill。

Enterprise AI platform for mapping Processes, Roles & Skills to identify AI transformation opportunities, skill gaps, and future workforce needs.

LLM-Powered Talent Intelligence Engine. An enterprise-grade ATS emulator utilizing hybrid matching (Exact, Fuzzy, Semantic) and Gemini 2.5 Flash to bridge the gap between candidate experience and job requirements.

India Runs — intelligent candidate discovery & ranking engine, built for the Redrob AI hackathon.

Open-source UK labour market research tool that turns ONS, Nomis and other official workforce, job advertising and earnings data into evidence for People and Talent decisions.

AI-powered candidate discovery and ranking system that evaluates skills, experience, recruiter engagement, and behavioral signals to identify the best candidates for a given job description.

Executive talent intelligence skill & plugin for Gemini Enterprise and Antigravity. Automated public web sourcing, verification, and structured C-suite benchmarking reports.

🧠 HireGraph — AI-powered multi-signal candidate ranking engine | INDIA.RUNS Hackathon 2026 | Beats keyword filters using semantic embeddings + career trajectory + behavioral intelligence

CPU-only pipeline that ranks 100,000 candidates in under 18 seconds, filters out fake profiles/honeypots, and generates factual recruiter justifications with DNA fingerprints & trust scores.

AI-powered Candidate & Talent Intelligence platform using Streamlit, LangChain/LangGraph, SQLite or Supabase, and Qdrant vector search.

A high-performance two-stage candidate ranking system for the Redrob AI challenge. Retrieves the top 100 fits from a 100k database, featuring honeypot excluders, skill-trust log-multipliers, and dynamic reasoning generation on CPU in under 10 seconds.