Resume

'Sup · Remote

• —Engineered a web scraping system using Python and Selenium to extract data from 30,000+ LinkedIn profiles, boosting targeted outreach efficiency by 40%
• —Developed an AI-driven cold email automation pipeline with Next.js, increasing engagement rates by 35% and cutting manual outreach time by 50%

UT Dallas · Richardson, TX

• —Designed and executed large-scale experiments on context length scaling for LLMs
• —Achieved 2x faster convergence and 50% lower compute vs. fixed 24k training while improving benchmark accuracy (+6% on AIME/AMC/MATH-500)
• —Introduced an iterative curriculum for context scaling for more efficient long-context reasoning and token utilization

XNode.AI · Remote

• —Developed a Neo4j knowledge graph with vector embeddings, centralizing company data (product, GitHub, specs)
• —Built an agentic chatbot with LLM and RAG for knowledge graph interaction, enabling queries and insights
• —Boosted knowledge graph query accuracy by 75% via RAG implementation