Open to opportunities

Ranjit Panda

@ranjitpanda

Senior Data Scientist and AI Architect building production-grade agentic AI and advanced RAG systems that drive measurable revenue.

India

Message

What I'm looking for

I’m looking to lead end-to-end production agentic AI and LLM/RAG systems—optimizing retrieval precision, latency, and cost—while owning scalable architectures across Azure/GCP/Databricks and partnering with product teams to deliver measurable business impact.

I’m an accomplished Senior Data Scientist and AI Architect with 4+ years of end-to-end expertise in engineering, optimizing, and deploying enterprise-grade Artificial Intelligence and Machine Learning solutions. I specialize in building robust Agentic AI frameworks—multi-agent systems and production-scale Advanced RAG applications—turning experimental research into high-ROI products.

At Tiger Analytics, I architect and engineer production-grade Agentic AI autonomous workflows using LangGraph and CrewAI, including deterministic state-management graphs for multi-agent execution and cross-validation. I also lead state-of-the-art Agentic RAG systems with Pinecone/Qdrant and techniques like query rewriting, hierarchical chunking, metadata filtering, and semantic re-ranking to achieve 98%+ contextual retrieval precision, plus PEFT/LoRA/QLoRA fine-tuning to reduce latency and infrastructure costs. Previously at A.P. Moller - Maersk, I delivered an inland transport dynamic pricing prediction engine that drove an $8M increase in corporate revenue, recovered over $10M by predicting booking cancellation risk, and architected a customer-query chatbot with response times under a second—while running the full MLOps lifecycle on Microsoft Azure.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist, AI Innovation

Current

Tiger Analytics

Jan 2024 - Present (2 years 6 months)

Architected enterprise-scale data ingestion and processing pipelines on Microsoft Databricks and PySpark to produce LLM-ready inputs. Built production agentic AI workflows and advanced agentic RAG systems using LangGraph/CrewAI and vector databases, and developed a multi-agent product/booking bot deployed on GCP.

Databricks Pyspark LangGraph Crew AI Agentic AI Pinecone Qdrant Query Rewriting PEFT QLoRA Mistral LLaMA GCP

Data Scientist, Logistics AI

A.P. Moller - Maersk

Jan 2021 - Jan 2024 (3 years)

Owned the end-to-end machine learning lifecycle on Microsoft Azure, including ETL-to-deployment pipelines and continuous drift monitoring. Developed a dynamic pricing prediction engine driving an $8M corporate revenue increase, cancellation-risk and customer-query chatbot solutions, and standardized hyperparameter optimization using Keras-Tuner.

Microsoft Azure MLOps ETL Drift Monitoring Dynamic Pricing Deep Learning Transfer Learning Chatbot Semantic Vector Indexing Keras Tuner

Analyst Intern

Deloitte Consulting

Jan 2020 - Jan 2021 (1 year)

Performed data preprocessing, feature selection, and exploratory data analysis to support executive decision-making. Improved predictive model precision/recall by 10%, designed A/B testing frameworks, built an ANN using Keras/TensorFlow, and researched LLM fine-tuning using LlamaIndex and a fine-tuned Mistral-7B for HR PDF question parsing.

data preprocessing Feature Engineering Statistical Validation A B Testing Keras TensorFlow LLM Fine tuning LlamaIndex Mistral HR Document Parsing

Computer Vision Intern

Digital Jalebi

Jan 2020 - Jan 2020 (0 months)

Curated, preprocessed, and augmented complex image datasets to improve training stability for deep vision models. Implemented and trained CNNs from scratch with hyperparameter tuning to improve accuracy on spatial classification tasks.

Hyperparameter Tuning