Basanite provides a two‑round interview platform that assesses software engineers on eight metacognitive dimensions, combining a CV‑driven conversational interview with a sandboxed coding session where candidates work alongside their chosen AI assistant. The system records interactions and extracts quote‑grounded evidence to generate detailed reports, enabling hiring teams to replace multiple screening rounds with a single, evidence‑based evaluation of genuine engineering capability.
Funding
Funding not disclosed
Founders
Product
Problem
Technical interviews often rely on isolated coding puzzles or static questionnaires that fail to assess a candidate’s real ability to perform engineering work, especially in AI‑augmented environments. This leads to hiring decisions based on superficial signals and creates a cheating arms race with AI tools, while overlooking crucial metacognitive skills such as judgment under ambiguity, tacit‑knowledge articulation, and effective human‑AI collaboration.
Solution
Basanite delivers a two‑round, construct‑templated adaptive interviewing platform that evaluates both what candidates think and what they actually do. In the first conversational round, AI‑driven, CV‑grounded questions probe eight metacognitive dimensions derived from cognitive science and human‑AI collaboration research, producing quote‑grounded scores. The second round places candidates in a sandboxed VS Code environment with a realistic codebase and a real ticket, allowing them to work alongside their choice of AI coding assistant while the system records keystrokes, prompts, and version‑control actions. The combined assessment generates evidence‑based reports that highlight genuine engineering capability, enabling hiring teams to replace multiple screening rounds with a single, rigorous evaluation.
Target Audience
Primary customers are engineering hiring teams at technology companies seeking to assess senior and mid‑level software engineers, AI specialists, and security engineers, as well as recruiting firms that need a rigorous, AI‑aware technical interview solution.
Features
- Construct‑Templated Adaptive Interviewing (CTAI) that asks unique, CV‑derived questions while using identical underlying constructs and scoring rubrics
- Evaluation of eight metacognitive dimensions: judgment under ambiguity, tacit‑knowledge articulation, intuition under data scarcity, psychological safety, creative problem reframing, ethical reasoning, transformative learning, and human‑AI collaboration intelligence
- Two‑round workflow: a 20‑30 minute structured conversational interview followed by a timed AI Collaboration Workbench coding session in a realistic multi‑thousand‑line codebase
- Tool‑agnostic AI assistance: candidates can use Claude Code, Cursor, Copilot, Aider, or any local CLI agent, with full instrumentation of prompts, keystrokes, and git activity
- Quote‑grounded scoring that extracts specific candidate statements as evidence rather than opaque numeric grades
- Automated identity verification and behavioural biometrics to ensure session integrity without restricting AI use
- Hiring‑team briefing reports that summarize dimension scores, key quotes, and recommended focus areas for final interviews