I'm a Data Science Manager at a leading US insurance firm, where I build production LLM pipelines, realtime fraud detection systems, and causal inference engines that drive millions in risk mitigation. I lead technical teams on building AI powered decision intelligence systems that actually work in production, not just in notebooks.

My technical identity bridges rigorous statistical foundations with modern LLM engineering. I believe in AI powered decision intelligence, the intersection of causal reasoning, Bayesian thinking, and production-grade orchestration. Epistemic rigor matters: every model claim should come with calibrated confidence.

I'm actively building: open source Bayesian network research libraries, public writing on production ML systems, and leading teams toward a vision of trustworthy, interpretable AI at enterprise scale. Moving toward Director/VP of AI & Decision Intelligence.

12+ years in data science and AI engineering

Philosophy

How I Think

Three core principles that guide my approach to AI and ML systems.

Epistemic Rigor First

Every model claim should have calibrated confidence. Uncertainty quantification isn't a nice-to-have—it's foundational to trustworthy AI.

Systems Over Models

ML value lives in end-to-end pipelines, not isolated notebooks. Architecture, orchestration, and monitoring matter as much as model performance.

Lead With Curiosity

Stay close to first principles even at leadership level. The best technical leaders keep their hands dirty with actual systems.

Timeline

My Journey

From statistical foundations to production decision intelligence.

2008

Started School in Computer Science

Fell in love with algorithms, data structures, and the power of code to solve complex problems. Built a strong foundation in programming and software engineering.

2012

First Job in working with Large Datasets

Built ETL pipelines and data warehouses for a major financial company, learning the importance of data quality, scalability, and real-world constraints in production systems.

2015

Back to School for Master in CS

Learned rigorous statistical inference, algorithms, and machine learning theory at Indian Statistical Institute. This mathematical grounding shaped my approach to production systems—when you understand the theory deeply, you make better engineering trade-offs.

2017

Large Scale ML in Production - Cybersecurity

Built single-handedly production ML pipelines for real-time cyber threat detection for a startup. This was my first deep dive into the challenges of deploying ML at scale—data drift, latency constraints, and the importance of monitoring and feedback loops.

2020

ML in HR Tech - Leading AI for Talent Acquisition

Led the development of AI-powered talent acquisition systems, leveraging machine learning to optimize hiring processes and improve candidate matching.

2024 - Now

AI in Insurance - Building Decision Intelligence Systems

Building production LLM pipelines, real-time fraud detection systems, and causal inference engines that drive millions in risk mitigation. Leading technical teams on building AI-powered decision intelligence systems that actually work in production—not just in notebooks.

Stack

Tools & Stack

Technologies I work with daily to build production AI systems.

PythonC++AWS BedrockPyTorchApache SparkKafkaLangChainFastAPIReactTypeScriptPostgreSQLDocker

Education

Master's Degree, Computer Science

Indian Statistical Institute (ISI), Kolkata

Foundational training in machine learning, statistical inference, and algorithms. This mathematical grounding shaped my approach to production systems—when you understand the theory deeply, you make better engineering trade-offs.