Quick facts
Every Personify student works one-on-one with a specialist like Anastasiia — not a generalist tutor.
Anastasiia Uvarova
Tech & AI Mentor — LLMs, NLP & Computer Vision (MIT)
Ana is a Master's student in Computer Science at MIT, completing her MEng after finishing her BS in Computer Science and Physics at the same institution. Her background blends deep theoretical grounding with hands-on AI research across NLP, LLMs, computer vision, and emergent communication.
Her most substantial work has unfolded over three years as a Research Assistant at MIT CSAIL, one of the world's premier AI labs, where she leads a study into symbol formation and emergent communication using behavioral data from 48 volunteer pairs. She built a Python Hidden Markov Model from scratch and engineered an interactive annotation system that improved data throughput up to 15×.
Her project work spans LLM security (evaluating the SecureLLM framework across LoRA, AdaLoRA, and (IA)3), NLP (extending RoPE embeddings with 2D character positions, pushing ASCII-table accuracy to 98%), and computer vision (3D street-structure prediction via point-cloud reconstruction).
Ana's stack includes Python, HMMs, Transformer architectures, PEFT methods, computer vision pipelines, and AWS. She's an exceptional mentor for students going deep on LLMs and fine-tuning, NLP research, computer vision, or AI research methodology.
Track Record
Research Assistant, MIT CSAIL (3 years)
Emergent-communication research; built an annotation system improving throughput up to 15×.
LLM security research
Evaluated SecureLLM across PEFT methods within 10% of a fully generalized model.
98% accuracy on structured text generation
Extended RoPE embeddings with 2D character positions.