Profile banner
Prabakaran Chandran

Prabakaran Chandran

Data Science & ML Leader, MS@Columbia, Decision Engineering Advocate

New York, New York, United States
pracha
Greater Bengaluru Area, India
weekday
Joined May 2026
Favicon imagepracha.me
Favicon imageprabakaranc98.github.io
Favicon imagegithub.com
Favicon imageprabakaranc.medium.com
Favicon imagesessionize.com

Network

18.3K connections
๐Ÿ“Š
Mu Sigma Network
๐Ÿข
Enterprise Data Analytics
๐Ÿค–
AI ML Specialists
๐Ÿค
Analytics Consulting Leaders
๐Ÿ’ป
Big Tech Data AI
๐Ÿ“ˆ
Product Growth Analytics
โš•๏ธ
Healthcare Pharma Analytics

Summary

Prabakaran Chandran is a highly experienced Machine Learning Engineer and Data Scientist with over six years in designing and deploying AI/ML solutions across diverse domains, including CPG, Oil & Gas, Power, Sustainable Aquaculture, and Data Observability. He possesses deep expertise in deep learning, causal inference, and graph machine learning, with a consistent focus on advancing decision-making systems and causal reasoning for complex enterprise and social challenges. sessionize+1
He is currently pursuing a Master's in Data Science at Columbia University, intending to deepen his theoretical understanding of Causal Inference, Probabilistic Modeling, and Reinforcement Learning. His academic pursuit aims to combine practical experience with rigorous scientific principles to architect future intelligent systems, demonstrating a commitment to continuous learning and foundational knowledge. pracha
Prabakaran has made significant contributions in various industry roles. At Informatica, he engineered a multi-agent AI system to automate troubleshooting of cloud data integration errors, utilizing LangGraph, RAG, and fine-tuned LLMs (like LLaMA and Phi3). At Captain Fresh, he developed computer vision models (YOLOv8, SAM) for aquaculture monitoring from satellite imagery and deployed MLOps pipelines on AWS SageMaker, improving data-driven decisions for farmers. His early career at Mu Sigma involved extensive project work for Fortune 500 clients, spanning supply chain transformation, demand sensing, agent-based modeling, and advanced computer vision and NLP solutions. pracha+2
Prabakaran is a dedicated educator and community builder in the AI space. He has taught over 300 hours of AI and analytics courses, trained government employees, led AI Bootcamps, and founded 'AI TamilNadu,' a non-profit dedicated to fostering AI education for rural students. He is a frequent speaker at Google Developer communities, TensorFlow user groups, and AI conferences, showcasing his commitment to sharing knowledge and promoting AI literacy. sessionize
His thought leadership extends to research and innovation, with publications on NLP, sentiment analysis, and causal AI applications. He is actively involved in curating reasoning datasets and training domain-specific LLMs for legal and insurance applications, reflecting a forward-thinking approach to real-world AI challenges. sessionize
Prabakaran has a unique perspective rooted in his Bachelor's degree in Instrumentation and Control Engineering, which he leverages to approach data science problems through a 'control theory' lens. This background informs his 'Decision Engineering' philosophy, emphasizing building frameworks for scalable, trustworthy decisions under uncertainty, and connecting data science challenges to systems thinking. pracha+1

Work

Education

Projects

Writing

Data Science Via Control Theory

June 1, 2025

Explores the connection between control systems (his undergraduate background) and data science, arguing that many data science problems can be viewed through a control systems lens. Discusses how this perspective helps in building intuition and problem-solving muscles, moving from 'sensors to tensors'.

Favicon imageprabakaranc.medium.com

Satellite Image Processing and Analysis using Python

July 1, 2023

A proposal for PyCon India 2023 outlining a comprehensive system for satellite image processing and analysis using Python. It covers data acquisition, preprocessing, image enhancement, feature extraction, and classification for diverse applications like agriculture and urban planning.

Favicon imagein.pycon.org

Cracking Data Science Interviews - Sharing My learnings from 25+ Interviews

December 1, 2022

Shares a framework for cracking data science interviews based on his experience, including self-assessment, grabbing attention (resume, job boards, LinkedIn), and continuous learning (Learn-ReLearn-UnLearn). Emphasizes the importance of hands-on delivery experience in advanced data analysis, ML, and deep learning.

Favicon imageprabakaranc.medium.com

The Most Effective Way to Learn Data Science!

December 1, 2021

Provides guidance on efficient and impactful data science learning, covering curriculum planning, learning support systems (communities), practicing environments, and a structured approach to tracking progress. Advocates for 'Data Science + X' to build unique projects and 'Learning in Public + Building in Public'.

Favicon imageprabakaranc.medium.com

Hobbies

Prabakaran is passionate about music and writing. wordpress+1