Skills: Large Language Models (LLM) | SQL | Text-to-SQL | LangChain | NLP | Fine-tuning | Python3 | Text-to-SQL
- Designed and deployed an end-to-end SQL co-pilot chatbot by fine-tuning the SQLCoder text-to-SQL model, leveraging extensive Retrieval-Augmented Generation (RAG) and prompt engineering to translate user queries into SQLite commands, execute them, and deliver clean, actionable results.
- Achieved 92% accuracy in answering financial domain-specific questions (up from 80%) by integrating Named Entity Recognition (NER) and a feedback loop, reducing costs and increasing savings up to 45% by eliminating the need for manual execution of complex SQL queries for routine tasks.

Skills: Natural Language Processing | Retrieval-Augmented Generation (RAG) | Python3 | PyTorch
- Engineered a Retrieval-Augmented Generation (RAG) system using large language models (LLMs) to perform semantic analysis of application resumes, accurately extracting and summarizing key achievements of applicants with the ECE department at NCSU.
- Executed extensive data preprocessing and annotation pipelines, leveraging natural language processing techniques to prepare training datasets. Fine-tuning pre-trained LLaMA-3.1-8B to specialize in educational document analysis.
- Conducted rigorous evaluation of the LLM-enhanced RAG system . Documenting the entire development lifecycle, including model architecture, training parameters, and performance outcomes, to provide comprehensive support for the ECE department at NCSU in their application review process.

Skills: Biomedical Signal Processing | Deep Learning | Python3 | Healthcare | PyTorch | SciPy | sqlite3
- Developed and integrated biomedical signal processing pipeline for cuff-less blood pressure estimation using ECG signals. Implemented filtering, segmentation, hand-crafted feature extraction, data augmentation on biomedical signals.
- Developed Deep Learning algorithms using LSTMs and Transformer technologies to automate blood pressure estimation for deployment on mobile edge-devices to facilitate real-time prediction on low powered edge-devices.
- Recipient of the Graduate Student Support Plan (GSSP), a highly competitive support package providing standard tuition coverage, in recognition of academic excellence.

Skills: Data Analytics |Customer Analytics | Python Developer | SQL | Database Management | pandas | scikit-learn
- Conducted data extraction from various sources, performing comprehensive cleaning, transformation, and aggregation of customer and sales data for savories and snacks to ensure high-quality datasets for developing analytics models.
- Implemented a clustering-based approach using DBSCAN to decompose the customer behavior prediction task, developing tailored models for each cluster that improved prediction accuracy and processing speed for large datasets.
- Developed and optimized a COWRF (COA-optimized Weighted Random Forest) model, achieving a 39.17% increase in processing speed and a 97.2% accuracy rate, marking a 4.7% improvement over previous models to evaluate the impact of promotional activities for snacks and savories, enhancing marketing strategy effectiveness for the products.

Skills: Biomedical Signal Processing | Machine Learning | Python3 | Healthcare | PyTorch | SciPy
- Developed and deployed a neonatal seizure detection system utilizing a scalable machine learning architecture based on ProtoNN, achieving high sensitivity and compact model deployment optimized for ultra-edge devices.
- Achieved a sensitivity of 87% and inference time of 243.92 milliseconds with a model size as small as 4.84 KB, ensuring rapid and accurate seizure detection on resource-constrained devices.
- Presented the research outcomes titled "Scalable Machine Learning Architecture for Neonatal Seizure Detection on Ultra-Edge Devices” at the Second International Conference on Artificial Intelligence and Signal Processing (AISP 2022), in collaboration with IEEE.

- Masters in Electrical Engineering: GPA: 3.38/4.0
- Academic Achievements: Recipient of the Graduate Student Support Plan (GSSP), a highly competitive support package providing standard tuition coverage for one semester, in recognition of academic excellence.
- Relevant Coursework: Neural Networks, Topics in Data Science, Automated Learning and Data Analysis, Digital Imaging Systems, Internet Protocols, Cloud Computing, Computer Vision

- Bachelor of Engineering in Electrical and Electronics Engineering; GPA: 8.85/10.0
- Relevant Coursework: Object Oriented Programming, Neural Networks, Machine Learning and Applications, Discrete Time Signals and Systems