Aravind Bandipelli

About Aravind Bandipelli

AI engineering and machine learning are my focus. I specialize in LLM reasoning evaluation, model behavior analysis, and prompt engineering — breaking down hallucinations, logic gaps, numerical errors, and domain-specific failures across coding, analytics, and finance tasks.I build evaluation pipelines, scoring systems, and benchmark tests that reveal how models reason and where they fail. My work strengthens accuracy, reliability, and alignment for production-ready AI systems.On the engineering side, I build with Python, React.NET Core, SQL Server, and Azure. Earlier, I contributed to Microsoft’s SPOC finance platform, improved SQL performance by about 70 percent, automated workflows, and delivered APIs, microservices, and analytics tools used by more than a thousand global finance users.I’m also building practical AI/ML projects — an LLM reasoning evaluation pipeline, a financial forecasting model, and an AI-generated React UI component library — all active on GitHub.Open to AI Engineer, Machine Learning, LLM Evaluation, and Applied AI roles across the U.S.

Experience

1. Ai Engineer

   Handshake

   Nov 2025 — Present · San Francisco, CA, US

   Evaluate LLM outputs on reasoning, coding, analytics, and financial tasks to improve model accuracy and reliabilityDiagnose reasoning failures, hallucinations, numerical errors, and bias patterns for model improvementDesign stress-test prompts to benchmark behavior and expose edge-case reasoning problemsPerform high-precision annotation, ranking, and scoring under strict AHT and reviewer-agreement standardsConduct finance-domain evaluations including numerical checks, statement interpretation, and regulatory reasoningBuild and refine React components using AI coding modelsDeliver evaluation summaries and analyses that guide model updates and dataset refinement

Education

• Indiana Wesleyan University

  Master of Science - MS

• Christu Jyothi Institute of Technology and Science

  Bachelor of Technology - BTech

  2018 — 2021