Open to opportunities

surajit dey

@surajitdey

Senior AI/ML Engineer building production GenAI, RAG, and multimodal pipelines for document understanding and compliance.

India

Message

What I'm looking for

I’m looking to build production-grade GenAI and RAG systems that combine text and vision, run reliably in the cloud, and improve document workflows end-to-end—especially agentic extraction and validation for real compliance use cases.

I’m a Senior AI/ML Engineer specializing in Generative AI, LLMs, and multimodal systems. I build production-grade pipelines for document understanding using RAG, agentic workflows, and transformer-based models—combining text and vision to improve robustness across real-world layouts.

In healthcare work, I designed and deployed end-to-end systems from PDF ingestion through classification, OCR, embedding, retrieval, and LLM extraction, including an agentic RAG workflow for clinical quality review. Across domains like legal and healthcare, I’ve led LLM-driven NLP pipelines, fine-tuned transformer models for extraction accuracy, and deployed scalable ML services using cloud infrastructure and production tooling.

Experience

Work history, roles, and key accomplishments

Senior Data Scientist

Foundation AI

Aug 2022 - May 2025 (2 years 9 months)

Built LLM-driven NLP pipelines for document classification and information extraction across legal and healthcare domains. Implemented RAG systems, fine-tuned transformer models for extraction, and deployed scalable ML using Ray Serve with CI/CD via GitHub and Jenkins.

LLM NLP RAG Transformer Models BERT PyTorch CI CD GitHub Jenkins

Data Scientist

KareerMatrix

Oct 2021 - Aug 2022 (10 months)

Designed and implemented a recommendation engine for the Hello Genius Child app based on user behavior and trends. Integrated it with AWS OpenSearch and deployed an “Insights” feature using NLP-based unsupervised classification and visual analytics.

Recommendation Systems NLP Unsupervised Learning OpenSearch AWS

Machine Learning Engineer

Wipro Infotech

Aug 2019 - Sep 2021 (2 years 1 month)

Built time-series forecasting models (ARIMA/SARIMA) for telecom data. Developed model validation frameworks to improve prediction reliability.

Time Series Modeling ARIMA SARIMA Model Validation

Senior AI/ML Engineer

Evernorth Health Services

May 2025 - Present (1 year 2 months)

Built a production-grade multimodal GenAI system for healthcare documents, enabling automated classification and structured extraction. Developed an agentic RAG system for clinical quality review with multi-step retrieval and validation, improving compliance detection and reducing audit time.