CV

Education

• Ph.D. in Computer Science and Applications, Virginia Tech, Blacksburg, VA, GPA: 4.00 (Expected: 2028)
  Research Focus: Interpretable and high-fidelity generative AI at the intersection of NLP and Computer Vision, leveraging Large Language Models and Diffusion Models
• M.S. in Data Science, Northeastern University, Boston, MA, GPA: 3.75 (2019 - 2021)
• B.E. in Electrical and Electronics Engineering, SJCE, Mysuru, India, GPA: 3.85 (2014 - 2018)

Professional Experience

• Data Scientist - Generative AI, iLink Digital (Jan - Aug 2024)
  • Designed and deployed a scalable Retrieval-Augmented Generation (RAG) system to process over 15M+ documents in under 0.5 seconds using Llama 2, Nvidia TensorRT, and CosmosDB for real-time decision-making.
  • Enhanced retrieval pipelines with hybrid search algorithms and reranking methods, significantly improving context relevance and minimizing hallucinations, validated through LLMEval and RAGAs.
  • Developed and launched a GraphRAG-powered chatbot and pet health information dashboard, processing over 6K+ documents via Neo4j for real-time entity extraction and historical insights.
  • Implemented adaptive retraining with live feedback loops, improving chatbot precision by 15% and recall by 12%, supporting diagnostic decisions for veterinarians.
• Data Scientist, Fidelity Investments (Feb 2022 - Oct 2023)
  • Built end-to-end NLP-based systems to extract, classify, and analyze financial data from reports using OCR, YOLOv5, GNNs, and Transformers, saving over 1000+ hours of manual labor.
  • Created a scalable recommendation system and anomaly detection pipeline to manage portfolios worth over $850M, generating additional revenue.
  • Fine-tuned and deployed LLM applications for code translation, summarization, and chatbot functionalities, leveraging PEFT, text-generation-inference, and agent frameworks.
  • Led and mentored a team of interns, guiding them in LLM research, solution design, and compliance-focused AI development, presenting findings to senior management.
  • Researched AI advancements and automated business processes as part of the Asset Management, evaluating solutions with custom KPIs and dashboards.
• Data Science Co-Op, Fidelity Investments (Jan - Jun 2021)
  • Forecasted stock prices for efficient portfolio management using advanced multi-variate time-series models resulting in returns beating the market. Conducted A/B tests for feature selection and built interactive real-time dashboards.
  • Built a large-scale distributed collaborative filtering-based pipeline for similarity mapping of billions of financial records using advanced SQL and PySpark. Tuned hyperparameters and set up auto monitoring-validation ML systems.
  • Simulated beta of assets by training and tuning an XGBoost regressor model on data with thousands of features using Map-Reduce. Speeded up execution by 15 times by reducing dimensionality and saved 40h/month.

Research Experience

• Graduate Research Assistant, Virginia Tech, Blacksburg, VA (Aug 2024 - Present)
  • Developing novel frameworks for high-fidelity domain-specific text-to-image generative tasks at the intersection of NLP and vision, by leveraging advancements in Large Language Models and Diffusion Models.
  • Researching methodologies to enhance domain-specific image generation capabilities of Diffusion Models by augmenting them with the instructional following capabilities of LLMs and innovative prompt engineering strategies.
  • Exploring sophisticated techniques to enhance interpretability of complex agentic frameworks for coherent and efficient task completion.
• Research Assistant, SJCE, Mysuru, India (Aug 2017 – Apr 2018) (Advisor: Dr. Neethi M)
  • Studied the effects of signal variation on the performance and longevity of electrical components by conducting extensive experiments in carefully simulated environments.
  • Designed novel frameworks that synergize machine learning and high voltage systems to automate and enhance the safety monitoring of complex, large-scale electrical circuits.
• Summer Research Fellow, IIT, Hyderabad, India (May – Aug 2017) (Advisor: Dr. Sumohana Channappayya)
  • Optimized hybrid solar-wind power generators by collecting large streaming data from numerical relays, using Deep Neural Networks. Boosted output voltage by 27%, beating existing benchmarks.
  • Collaborated with chemists and computer scientists to develop cutting-edge machine learning and vision methodologies to identify precise drug molecular structures with minimal side effects.
  • Contributed to novel drug discovery efforts by synthesizing over 100K high-quality synthetic data samples with GANs.

Skills

• Machine Learning & AI: NLP, Generative AI, RAG systems, Transformers, Diffusion Models.
• Programming: Python, R, SQL, PyTorch, TensorFlow, Hugging Face.
• Deployment & Cloud: AWS, Azure, Docker, FastAPI, Streamlit, Triton Server.
• Visualization: Tableau, PowerBI, Matplotlib, Seaborn, Plotly Dash.

Publications

• Context Canvas:
  • Enhancing text-to-image diffusion models with knowledge graph-based RAG and self-correction.
• FluxSpace:
  • Domain-agnostic image editing using rectified flow transformers for semantic control.
• Fault Analysis and Predictive Maintenance (IEEE Xplore, 2018):
  • Pre-fault detection of electrical systems; awarded Best Paper at ICEECCOT-2018.

Talks

• Crafting Balance: Open-Source LLMs for Responsible AI (UMass Amherst, Mar 2024)
• Harnessing the Power of LLMs: Practical Solutions (Analytics Vidhya Webinar, Oct 2023)
• Generative AI Workshops (Northeastern University, 2023): Focused on AI deployment strategies and career development.

Teaching

• Graduate Teaching Assistant, Virginia Tech (2024 - Present):
  • Introduction to AI (CS4804).
• Graduate Teaching Assistant, Northeastern University (2020 - 2021):
  • Programming with Data (DS2000).

Service & Leadership

• Women in Data Science Ambassador, Greater Boston (Feb 2024 - Present).
• Mentor, Women in Big Data (May - Sep 2024).