Ananta Verma

I am an AI Engineer and Data Scientist passionate about building practical GenAI applications and machine learning systems that solve real-world problems. My work spans AI agents, retrieval-augmented generation (RAG), multimodal pipelines, and applied ML for industries from healthcare to e-commerce.

Recently, I

• Built a hallucination evaluation pipeline leveraging LLM judges, Langfuse tracing, and Labelbox review workflows to benchmark reliability of medical note generations.
• Developed an Agentic Scribe with CrewAI, an 8-agent pipeline parsing transcripts into clinical notes, reducing manual effort and automating compliance checks.
• Designed an ELT weather automation pipeline to stream, clean, and visualize real-time weather data for decision-making.
• Created a multimodal RAG system combining satellite imagery and historical text for the OpenAI-to-Z Kaggle Hackathon, orchestrating agents for hypothesis generation and evaluation.
• Led a Shopify data analysis project for Pashion Footwear, revealing SKU profitability insights that informed business strategy.

Beyond projects, I actively participate in Bay Area hackathons every weekend, where I prototype agentic systems, multimodal AI tools, and real-time ML pipelines, fueling both creativity and continuous learning.

Technical Strengths

Python, PyTorch, R, SQL, NLP, LLMs, RAG, multimodal AI, AI agents, data pipelines, computer vision, and cloud deployment (AWS, Docker).

Research
Published in IEEE on Early Detection of Cardiac Arrest Risk Factors, achieving 98% accuracy with Random Forest, outperforming KNN, Logistic Regression, and XGBoost.

I thrive in collaborative, impact-driven environments, especially where AI meets sustainability, healthcare, or public services. Currently seeking full-time opportunities as an AI/ML Engineer or GenAI-focused Data Scientist to keep building scalable, human-centered solutions.