Open to opportunities

Avani Patle

@avanipatle

Associate Data Scientist specializing in LLM-powered automation, RAG pipelines, and ML/GenAI systems for real-world decisions.

India

Message

What I'm looking for

I’m looking to grow as a Data Scientist building production-ready GenAI/LLM systems—RAG, evaluation, and monitoring—alongside teams that value measurable impact, clean engineering, and fast iteration from research to deployment.

I'm a Data Scientist specializing in production-grade agentic AI—LLM-powered automation, multi-agent architectures, and applied AI research for customer-facing and domain-specific workflows, with a strong interest in the intersection of AI and quantitative finance.

At ICICI Lombard, I architect and deploy production voicebots built on a thin-LLM + deterministic-graph agent design—orchestrating intent routing, tool use, and multi-journey conversational flows over RAG (FAISS) and STT/TTS pipelines. Containerized on Azure Container Apps with Pipecat, these systems handle high-volume customer support 24×7. I've driven latency and scaling optimizations (6s → 2–3s end-to-end) through asynchronous pipeline orchestration and concurrent tool execution, and I build the testing infrastructure—500+ automated test cases, conversational harnesses, intent validation—that keeps agents reliable in production. Alongside delivery, I lead applied research and benchmarking across LLMs, OCR, and multimodal models to guide production adoption across teams.

I'm also a published researcher (IEEE, first author) in NLP and medical imaging.

Experience

Work history, roles, and key accomplishments

Current

Associate Data Scientist

Current

ICICI Lombard

Aug 2025 - Present (1 year 1 month)

Architected and deployed an LLM-powered health insurance voicebot with STT/TTS and RAG over policy documents, enabling 24×7 automated policy support and reducing agent call load. Built a PED extraction pipeline to detect and timestamp pre-existing disease declarations from call transcripts, reducing manual review time and mitigating premium leakage risk.

LLM RAG Faiss Azure OpenAI Azure Container Apps Pipecat Data Pipelines

Quant Intern

Orim Advisors Private Limited

Jan 2025 - Jun 2025 (5 months)

Automated volatility and quarterly split computations for 5,000+ listed companies to improve signal accuracy and built a model mapping stock breakouts in 150+ F&O stocks to news events for event-driven analysis. Developed a deterministic keyword extraction pipeline for 500+ financial documents and deployed scalable FastAPI applications for high-frequency query processing.

Algorithmic Trading Volatility Financial Document Processing Keyword Extraction fastAPI Python Data Pipelines

Gen AI Research Intern

Cornerstone Ventures

Jul 2024 - Sep 2024 (2 months)

Benchmarked 10+ LLM models across capability, pricing, and latency and produced structured evaluation reports to guide enterprise deployment decisions. Conducted market research across 10+ SaaS companies in healthcare and enterprise AI, performing TAM and opportunity analysis to identify scalable generative AI applications.

LLM Benchmarking Model Evaluation Generative AI Market Research Enterprise AI

Machine Learning Intern

SequelString AI

Jun 2024 - Jul 2024 (1 month)

Developed invoice extraction solutions using OCR and machine learning to automate financial data parsing and validation. Built automated email workflows adopted by 15+ internal users to streamline communications.

Computer Vision Machine Learning Document Processing Data Validation Python