Dipali Patidar

About Dipali Patidar

I’m an Applied Machine Learning engineer who bridges data science and software to deliver production-ready ML solutions that move key business metrics. Across 8 years in cloud, gaming, and digital media, I’ve owned the full model lifecycle—from data pipeline design and feature engineering through experimentation, deployment, and post-launch monitoring.I pair solid statistical grounding with hands-on experience in NLP, recommendation, and generative-AI systems. My expertise lies in designing and implementing innovative solutions to complex problems across various industries. I have successfully designed and trained models for diverse NLP tasks, including text summarization, text classification, and language modeling. With a deep understanding of AI/ML algorithms and techniques, I am proficient in programming languages such as Python, Java, and SQL, and adept at using deep learning frameworks like Keras and PyTorch.I have contributed to team projects involving the fine-tuning, quantization, deployments, inferencing, and benchmarking of large language models (LLMs) in a collaborative environment.I am passionate about utilizing the latest advancements in AI, machine learning, and LLMs to create innovative solutions for real-world challenges. My experience includes working in cross-functional teams, collaborating with data scientists, software engineers, and product managers to deliver high-quality NLP solutions that align with business objectives.

Experience

1. Senior Machine Learning Engineer

   Oracle

   Aug 2023 — Present · Redwood City, CA, US

   Document Understanding Solution: Working on building pipeline to convert PDFs to Markdown/JSON using multimodal LLMs and layout parsing for enterprise knowledge bases.– Agent Evaluation Framework: Designed LangChain + LangGraph system to measure chatbot accuracy, latency, and hallucinations, accelerating release validation- LLM Infrastructure & Optimization: Built Dockerized vLLM + TGI stacks for scalable text/multimodal inference, then benchmarked, quantized, and tuned multi-GPU deployments to cut latency and memory without accuracy loss.

Education

• Indian Institute of Technology, Roorkee

  Bachelor of Technology (B.Tech.), Electronics and Communications Engineering

• University of Florida

  Master's degree, Computer Science