At the Microbial Biophysics & Design Lab, IIT Delhi, I processed 50+ datasets and 10,000+ data points across four Chilika Lake zones. I extracted water-quality and remote-sensing indices from satellite imagery and used principal component analysis to identify key variance drivers.
I benchmarked Multiple Linear Regression and Random Forest models for prediction and bacterial-load mapping. My work used Python, Pandas, Google Earth Engine, QGIS, SciPy, and Scikit-learn.
On my Speech Recognition & Speaker Diarization Pipeline project, I’m developing a deep-learning pipeline with pretrained transformer models. I’m implementing LoRA and PEFT fine-tuning and building a FastAPI inference service while exploring vLLM model serving.
I also built projects in tweet analysis, credit-risk prediction, and C++ systems. These included an XGBoost model for tweet engagement, a credit-default prediction pipeline with SHAP explanations, and an in-memory key-value store.

