Teaching Assistant, Automata, Formal Languages & Computability
Office hours and grading for a graduate automata theory course. Helping students work through formal proofs is harder to teach than it looks.
TeachingTheoryPython
Available — ML engineering
Computer VisionMLOpsLLMsFederated ML
Hey,I'mKaavin.
CS grad student at Rice working on ML engineering. I focus on the deployment side of things: fine-tuning, monitoring, and getting models to work reliably outside of notebooks. TAd for automata theory last semester.
LLM Fine-TuningMLOpsComputer VisionAWSFederated LearningModel MonitoringApplied MLLLM Fine-TuningMLOpsComputer VisionAWSFederated LearningModel MonitoringApplied ML
Projects
Featured
LLaMA Fine-Tuning for Research Classification
Fine-tuned LLaMA with LoRA on arXiv papers for research topic classification. Accuracy went from 40% to 67%, mostly through better data curation.
PyTorchHugging FaceLoRA
FeaturedMLOps Sentiment Pipeline on AWS
End-to-end sentiment analysis pipeline on AWS EKS. Data versioning with DVC, experiment tracking with MLflow, production monitoring with Prometheus and Grafana.
AWSDockerMLflowDVC
Federated Learning for Wind Prediction
Federated LSTM for wind prediction using Flower. Each node trains locally and only shares model updates, keeping raw data on-site.
TensorFlowFlowerLSTM
COVID-19 Detection from Chest X-rays
Compared a custom CNN and VGG-16 for COVID-19 detection from chest X-rays. VGG-16 outperformed significantly with less training data.
TensorFlowKerasCNN
Face Liveness Detection
CNN-based liveness detection to distinguish real faces from spoofs. Used heavy augmentation to handle lighting variation across different environments.
TensorFlowOpenCV
Realtime Spam Detection
Real-time spam classifier connected to IMAP. Logs feature distributions over time for drift detection.
scikit-learnTF-IDFIMAP
Experience
Software Engineering Intern
Built a face-recognition attendance system used by 500+ students on campus. Optimized inference to under 0.5s and shipped a companion React Native app that went to production.
Computer VisionReact NativeOpenCV
Machine Learning Research Intern
Built LSTM models for wind energy forecasting, improving on the baseline by around 10%. Also set up a federated training pipeline with Flower to keep raw data local across sites.
LSTMFederated LearningFlower
Technical Stack
Languages
PythonJavaRSQL
ML & Deep Learning
PyTorchTensorFlowKerasscikit-learnHugging FaceLoRALLaMAFlower
Data & Cloud
AWS (EKS, EC2, S3)DockerMLflowPrometheusGrafanaMongoDBDVC
Tools & Frameworks
OpenCVPandasNumPyMatplotlibReact NativeGitCI/CD
Publications
AI-Powered Attendance System Using Facial Recognition
Published from the VIT internship. Covers the system architecture and deployment decisions, including latency optimization and inference placement.
Enhancing Drug Repositioning Through Collaborative Metric Learning
Collaborative metric learning for drug-disease association prediction with improved ranking performance on CTD benchmarks.
AI Applications in Nutrition & Education
Survey of applied AI in nutrition and education, with a focus on deployment context and practical evaluation beyond benchmark accuracy.
Certifications
NPTEL - CSRNPTEL - Educational LeadershipNPTEL - Emotional Intelligence
Looking for ML engineering or research roles.
Especially interested in MLOps and LLM systems, but open to most things. Feel free to reach out.
3
2
6+