Divyanshu Bhardwaj

• Architected and led the full-stack design of a multi-phase Autonomous AI Agent platform on AWS (using EC2, API Gateway, and DynamoDB) for high-volume system log analysis and automated Root Cause Analysis (RCA), adopted across multiple product lines. • Engineered the agent's multi-step reasoning framework, leveraging LLM Inference and proprietary knowledge integration to automate complex log parsing, substantially reducing technical investigation time. • Led technical strategy and mentorship, establishing production-grade CI/CD pipelines and engineering best practices to ensure scalable, documented deployment. • Developed and maintained critical documentation (including API specifications and model performance guides) for all deployed GenAI products, facilitating adoption and maintenance by engineering support teams.

• Designed the architecture and developed the core clustering model for an event analytics system, providing technical oversight to a junior team implementing the real-time MLOps pipeline (Kinesis, Lambda, S3) that improved incident detection. • Built domain-specific LLM agents using RAG and VectorDB, enhancing AI-based virtual assistance capabilities. • Collaborated with Customer Support & Product teams to engineer a comprehensive, automated Customer Satisfaction Score (CSAT) system, including a bias mitigation function, driving targeted operational improvements.

• Conducted text data mining on a large corpus of documents related to stopping modern slavery in the global supply chain, leveraging NLP and data pre-processing techniques such as tokenization, lemmatization to uncover insights and patterns • Created a custom data pipeline to extract PDFs & texts from 50k URLs, using multithreading to reduce processing time by 80% • Applied advanced NLP techniques such as Latent Dirichlet Allocation (LDA) to identify hidden semantic structures in the documents, revealing topics and themes related to modern slavery in the supply chain

INFO 6106 - Neural Network Methods and Tools ARTG 5320 - Intro to Statistics & Data Analysis INFO 5002 - Intro to Python • Assessed student’s learning to determine weaknesses & needs, and improved their coding skills through assignments • Supporting the professor to mentor and manage 30+ students’ performance reviews through dashboards

• Collaborated with cross-functional stakeholders to translate business requirements into technical ML specifications and identified high-impact data sources for model development. • Optimized incident-correlation algorithms for large datasets, improving system efficiency and reducing manual investigation effort by 40% across 2M+ records.

• Exercised EDA using data mining & web scraping to identify areas that required manual effort reduction across multiple instances • Planned, designed, and implemented a python-based automation for clients, resulting in $150K cost savings annually • Automated SOX team process using python code which saved manual efforts of 120 hours a week and 30% manpower reduction • Monitored KPIs and DelEx parameters, interfaced with clients to evaluate business needs & proposed convenient solutions • Developed complex SQL queries and used Python to extract reports from a large database with over 2 million rows of data, ensuring efficient data processing and retrieval • Deployed an auto ticket triaging tool which reduced 90% resource efforts that had to work on 24/7 shift • Collaborated with cross-functional team to implement processes that maintained SOX compliance • Well acquainted with Software Development Life Cycle (SDLC), Agile methodology