• Developed features and bug fixes in Python for ArangoDB’s official LangChain integration, reaching 50k downloads on PyPi.

• Introduced advanced similarity search strategies (Jaccard, Dot Product, Max Inner Product) in ArangoVector.

• Enhanced the GraphRAG importer service by introducing a Relationship Types parameter, enabling more precise representation of graph structures.

• Founding engineer of ArangoDB’s internal Knowledge Enablement tool using Python, GraphRAG techniques, and the Slack SDK, providing real-time knowledge enablement to +100 employees.

• Collected and analyzed 100+ state blockchain and AI statutes, 50+ legislative histories, and 20+ working group reports into a structured legal dataset.

• Performed NLP analysis on 150K+ words of statutes, supporting research on technology legislation.

• Built semantic tagging pipeline using LLMs and vector similarity to cluster legal provisions for better retrieval.

• Developed a multi-technology framework for methane emissions modeling, analyzing Probability of Detection (POD) across satellite, aerial, and ground-based sensors to improve emission estimates.

• Built an ETL pipeline to extract research papers, text, and images from 100+ documents, reducing processing time by ~90% and automating data retrieval.

• Leveraged LLaMA & Hugging Face for document summarization and prompt engineering, classifying 100+ research papers on methane emissions for structured insights.

• Fine-tuned VGG16 using PyTorch to classify methane-related graphs with 92% accuracy, training on 4,000 web-scraped and synthesized graphs.

• Developed a YOLO-based image-to-data extraction framework, integrating AWS Bedrock and a custom ChartOCR pipeline to digitize methane emissions graphs, targeting 93% precision and 94% recall for structured data extraction.

Worked on the "Literature Review and Applications of Generative AI" project under Prof. U Dinesh Kumar.

This paper aims to give an understanding of Generative AI, its history, Hidden Markov Models, Large Language Models, types of models, and applications of generative artificial intelligence in various domains and industries.

• Reviewed 50+ papers on Generative AI and LLMs, documenting its evolution and mathematical foundations as an intern.

• Analyzed 5+ foundational models – Compared architectures, performance, and applications in GenAI research.

• Analyzed transformative applications of Generative AI in 10+ industries, highlighting key models/products and their impact.