About
Building reliable AI systems and practical products with a clear engineering mindset.
I work across LLM applications, retrieval pipelines, medical imaging research, and data-driven software — turning strong models into tools that feel useful and well built.
Recent work
2026University of Maryland, College Park
Recent interests
OngoingDiving deeper into LLM inference
Exploring inference systems, cluster communication, and how modern serving stacks scale.
A stark look at the work — responsibilities, impact, and tools.
Oct 2025 — Dec 2025
Data Science Intern
Engineered automated RAG pipeline and optimized ML deployment for healthcare AI systems.
Oct 2023 — Jun 2024
Junior Machine Learning Engineer
Developed and deployed state-of-the-art ML models for various NLP tasks.
Feb 2023 — Sept 2023
Software Developer Intern
Converted monolithic architecture to multi-server and performed large-scale data extraction.
Fundamentals, applied machine learning, and work outside the classroom.
LLM inference, Model Serving, Optimization, and systems engineering projects.
01
Developing a diffusion-based model for generating healthy counterfactual brain MRI scans while preserving the patient's anatomy. Enables further analysis of pathological features by comparing real and generated scans.
02
Fine-tuned Microsoft CodeReviewer model on code refinement tasks and integrated it into GitHub Actions. Model metrics were logged on ClearML and the model was containerized and deployed on HuggingFace Spaces.
03
Solved a convex optimization problem using mathematical solvers to predict future financial trends with directional accuracy. Introduced wrong trade penalty terms often overlooked in financial market prediction.
04
Implemented custom Kruskal's algorithm to visualize information spread on Twitter dataset with 80k+ nodes. Used Graphistry for interactive visualization of graph-based information propagation.
05
Training an SVM to accurately predict the next word from newsgroup dataset corpus. Implementing 5-fold cross-validation to increase model accuracy while preventing overfitting.
06
Developing a reliable and accurate method for diagnosing Alzheimer's disease at an early stage using 3D MRI scans. Improving patient outcomes through earlier intervention and treatment.
07
Predicting students' chances of admission to their dream university using a linear regression model. Collecting and analyzing various scores necessary for the application process.
A slow-moving strip — hover to pause, open to view.
Open to conversations around machine learning engineering, AI products, and thoughtful collaboration.
