Key information
San Francisco-based startup Mirror Particle is developing an alternative approach to predicting human behavior by building a custom foundation model from scratch. Co-founder and CEO Abhivyakti Ahuja argues that current prediction methods relying on large language models are flawed because they model written language rather than visual perception, spatial reasoning, and social intelligence.
Unlike static approaches, the startup aims to capture longitudinal data showing how people change over time, tracking triggers and shifting motivations. The proprietary engine combines client customer data, pop culture, current events, and social media to focus heavily on actual revealed behavior rather than traditional self-reported survey answers.
Mirror Particle focuses its initial go-to-market strategy on market research alongside brand and product strategy, helping brands understand not only consumer preferences but also the underlying reasons behind them. For example, an early pilot with a pet food brand revealed that packaging imagery was less critical than overcoming a mass-market and cheap perception issue.
Ahuja co-founded the two-year-old company alongside Will Song and Thomson Yen after previously working at Amazon Robotics. The enterprise has secured an angel round, approaches its first venture round, and will compete in Startup Battlefield 200 at TechCrunch Disrupt 2026 in San Francisco.
