Designing and building eval training pipelines and agent behavior scoring rubrics for a production multi-agent AI platform. Focused on goal drift detection and failure mode classification. Contributing as a volunteer.

Building AI teaching assistants that combine large language models, retrieval-augmented generation (RAG), and autonomous agent architectures, while researching human-simulation agents for behavioral modeling and emergent interaction studies.

Backed by non-dilutive capital and working on a startup in the edtech space. Building innovative solutions to transform educational technology and learning experiences.

Developing models for dynamic role assignment and efficient communication through graph-based protocols. Working with Multi-Agent Reinforcement Learning.

Conducting research on causal inference and causal effect estimation using DoWhy libraries and causal models. Working under Principal Investigator Dr. Gabriel Terejanu to develop advanced methods for understanding causal relationships in complex systems and applying these techniques to real-world problems.

Researching predictive models and creating trends. Used for weather prediction and pollution management.

Assisted in teaching pre-calculus to high school students at the #1 public school in the USA. Provided one-on-one tutoring and helped develop curriculum materials.

Researched and built applications for Lead Poisoning Awareness.

Conducted research on large language models for scientific applications. Fine-tuned models with multiple prompt engineering techniques to analyze efficiency.