Abhinav Kumar

Hey there! I am Abhinav, ML lead at Teammates.ai. Most AI interview tools use LLMs to generate scores — we don’t. Language models are great at generating, but too random and opaque for high-stakes evaluation. Temperature, hallucinations, token bias — none of that belongs in candidate scoring. Instead, Sara uses structured feature extraction + predictive modeling. We benchmark performance...