Anushka Sharma

Anushka Sharma

she/her

Undergraduate AI researcher specializing in open-ended evolution and AGI safety. I design evolutionary frameworks that sustain novelty without explicit fitness functions, addressing reward hacking in AGI.

Anushka Sharma

I’m an undergraduate AI researcher and third-year B.Tech student at Banasthali Vidyapith, specializing in open-ended evolution and AGI safety through constraint-driven evolutionary systems. My work sits at the intersection of evolutionary computation, intrinsic motivation, and cognitive architectures — designing frameworks that sustain novelty generation without explicit fitness functions, directly addressing reward hacking in artificial general intelligence.

My primary project, the Genesis Engine, is a sham-controlled co-evolutionary platform I built independently to discover AI failure modes. It has produced statistically significant results across multiple versions and has led to submissions at GECCO 2026 and CCS 2026. I also published a book on AI for non-technical audiences at age 17.

Beyond research, I build applied AI systems — from edge AI navigation for visually impaired users (Qualcomm Edge AI Hackathon finalist) to embedded ML pipelines and autonomous robotics.

Research Interests

  • Evolutionary Computation & Open-Ended Evolution
  • AGI Safety & Reward Hacking
  • Intrinsic Motivation & Constraint-Based Learning
  • Cognitive Architectures & Multi-Agent Systems
  • Embedded Intelligence & Edge AI

Recent Highlights

  • GECCO 2026 (Under Review): Sustained Evolutionary Activity Without Fitness Functions — Genesis Engine V2, achieving 58.3% sustained activity rate over 10,000 generations (p < 0.01, Cohen’s d = 1.47)
  • CCS 2026 (Submitted): Abstract on structure/function decoupling as a precursor to deceptive alignment, based on Genesis Engine V5 findings
  • Qualcomm Edge AI Hackathon 2025: Top 10 finalist from 650+ teams nationwide with VisionIRL — offline AI navigation for visually impaired users
  • Google Girl Hackathon 2025: Advanced to Round 3 with SynthPredict — 98% accuracy ML model for RTL synthesis prediction

Genesis Engine Projects

Contact

📧 anushka.care@gmail.com    📱 +91 79062 08181    📍 Rajasthan, India

Available for remote internships from March 2026 (part-time) or June 2026 (full-time). Willing to relocate for conferences or extended collaboration.