Experiences

Driving AI-powered automation projects focused on intelligent document processing, LLM-based analytics, and workflow integration using tools like Power Automate, Power Apps, and Python. Projects include email-to-database pipelines, web scraping with real-time summarization, and interactive dashboard development.

• Developed predictive maintenance models by applying machine learning to industrial sensor data and SAP systems
• Built interactive dashboards to support failure analysis, enabling cost-effective maintenance strategies
• Designed a web-scraping system using Selenium and BeautifulSoup to extract brake-related news from industry sources
• Integrated Ollama for LLM-based summarization of scraped content and deployed results via interactive Python dashboards
• Built an end-to-end automation pipeline using Power Automate to parse incoming emails, extract structured data using AI, store it in SQL Server, trigger approvals via a Power Apps interface, and sync validated data to Excel and SAP

Conducted research focusing on binary hardening and vulnerability mitigation, working with cutting-edge tools to identify and patch security vulnerabilities.

• Performed binary exploitation and vulnerability mitigation using Ghidra, Patcherex2, and Dockerized PHP environments
• Identified and patched CVEs through manual reverse engineering, static and dynamic analysis
• Conducted heap overflow diagnostics and root-cause tracing
• Reproduced exploits on 32-bit and 64-bit architectures
• Leveraged AddressSanitizer and GDB for comprehensive vulnerability analysis
• Implemented source-level fixes with secure memory allocation strategies

Worked on an innovative agricultural AI project utilizing drone technology and artificial intelligence to revolutionize farming practices.

• Developed data models using Python and PostgreSQL to track weed types, herbicide usage, and historical weather data
• Designed a Tableau dashboard to visualize weed distribution and optimize herbicide application
• Collaborated with team to improve weed detection models, enhancing accuracy and reducing environmental impact

Mentored and guided first-year computer science students in fundamental programming tools and concepts.

• Conducted Practical Study Observations (PSO) sessions for 900 first-year CS students
• Provided instructions on Unix/Linux commands, terminal navigation and GitHub
• Graded assignments and engaged with students on Ed Discussion platform
• Clarified complex topics and enhanced learning outcomes through one-on-one support

Developed and optimized advanced forecasting models for mix forecasting and demand prediction, working on 2-stage hierarchical systems that significantly improved accuracy across multiple product levels.

• Developed, implemented, and optimized predictive models (XG-Boost & RandomForest) using advanced data processing
• Achieved a forecasting accuracy of 90% at the platform level and 70% at the product level (PLID)
• Drove development of a 2-stage hierarchical demand forecasting model
• Improved accuracy with a combination of multivariate and univariate ML algorithms
• Showcased results in a dynamic Python-based dashboard