Hi I'm Tejus, welcome to my site.

About me

Completing my Master's in CS at UC San Diego, specializing in Artificial Intelligence.
Research user-LLM representational alignment to enable LLMs to be extension of your own mind.
Passions lie in research-driven, open-ended problems in model behavior, applied ML, computational modeling.

Currently:

Conducting research on representational alignment between human and LLMs. Learn more about my research here.

Previously:

Quantitative Research Intern @ Evovest - Time-series Diffusion models for equities forecasting.
- Altered noise schedules, sampling distributions, and CFG to enable model to learn useful signal from extremely low SNR data.
- Demonstrated a significant improvement in the learned correlation structure and uncertainty forecasts.
- Quantified which part of the signal can only be captured by diffusion, compared with other transformer and neural network architectures.
Worked on an implementation of a quantized transformer on FPGA for low latency, at UCSD.
- Model classified short-lived particle streams generated by the Large Hadron Collider.
- Accuracy-preserving quantization and fast inference were a focus.
Researched physics-informed deep learning methods for volumetric medical imaging at UCI.
- Work accepted by the International Journal of Cardiovascular Imaging.

Resume