nthexperiment

my personal lab for agentic and decision sciences (AI and Humans) - a road to the destiny: decision engineering as a practice

👨‍🔬
About This Lab
The journey behind agentic decision science
Prabakaran Chandran, a dreamer and doer in the landscape of data science and machine learning; a first-gen graduate from an unknown village, chasing the north-star in the heart of big-apple. A control systems undergrad - seeing enterprise problems as systematically controllable and traceable via data and science.

Currently pursuing my self-designed graduate degree in "Computational Data Science and Decision Engineering" through which I am aiming to build experimental and creative things. I work as a practice lead at a stealth startup focused on multi-modal machine learning and document processing, while simultaneously exploring linear mixed effects models, predictive inference, and advancing through MOOCs on causal inference.

"Learning is the only mean that corrects our trajectory, and brings clarity to our chaos."

Over the past six years, I've been answering the questions that drive organizational decision-making: "What causes drift in petrochemical pricing?", "Why do certain demographics prefer specific brands?", "Can we predict tomorrow's energy output with confidence?"

My evolution from modeling industrial systems with MATLAB during undergraduate studies to leading teams in supply chain optimization, demand sensing, and agentic AI solutions has been driven by one persistent question: Why do we still default to association-centric approaches when P(y∣x) ≠ P(y∣do(x))?

This realization shifted my focus from prediction to understanding the "true why" behind variable interactions. Through projects spanning Fortune 500 companies to innovative startups—from building Bayesian networks for Coca-Cola's consumption patterns to developing physics-informed neural networks for manufacturing optimization—I've witnessed the transformative power of causal reasoning in real-world applications.

nthexperiment represents my commitment to advancing "decision engineering"—a next-generation approach that blends causality, uncertainty quantification, and adaptive systems to build enterprise-scale decision-making solutions at the intersection of machine learning and Causal AI.

🚀
Why Decision Engineering?
A balanced synthesis of technical rigor and systems thinking

🧠 Core AI/ML & Data Science Foundation

Decision engineering demands mastery of causal inference, uncertainty quantification, and predictive modeling. It's not enough to build models—we need to understand when P(y∣x) ≠ P(y∣do(x)) and design systems that can reason about interventions, not just correlations.

🎨 Design & Systems Thinking

Human-centered design meets complex systems architecture. Decision engineering requires understanding how humans actually make choices, designing interfaces that support better reasoning, and building scalable systems that adapt to organizational needs and cognitive limitations.

🔍 Skepticism & Empirical Research Mentality

Every assumption must be questioned, every model validated through rigorous experimentation. Decision engineering embraces intellectual humility—acknowledging uncertainty, designing for falsifiability, and continuously updating beliefs based on evidence rather than intuition.

🌐 Complexity Science Understanding

Business ecosystems are complex adaptive systems with emergent behaviors, feedback loops, and non-linear interactions. Understanding network effects, phase transitions, and system dynamics is crucial for predicting how interventions will cascade through organizations.

⚖️ The Integration Challenge

Traditional approaches silo these domains. Data scientists ignore human factors, designers underestimate technical constraints, researchers overlook system complexity. Decision engineering synthesizes all four pillars into coherent, actionable frameworks.

🎯 The Decision Engineering Synthesis

By balancing technical depth with systems thinking, empirical rigor with design empathy, and complexity awareness with practical implementation, decision engineering creates robust solutions that actually work in the messy reality of organizational life.

"Decision engineering is where rigorous science meets human reality—building systems that are both technically sound and humanly usable."

🧠
Core AI/ML & Data Science
Essential learning areas
Causal Inference & Agentic AI
Building AI systems that understand causality for nudging human and business decisions
Graph Neural Networks
Foundational models for relational data, graph embeddings, and network analysis
Reinforcement Learning
Decision-making algorithms, policy optimization, and multi-agent systems
Deep Learning & Foundation Models
Neural architectures, transformers, and representation learning techniques
Predictive Inference & Uncertainty
Modeling uncertainty, Bayesian methods, and uncertainty quantification
Time Series Forecasting
Temporal pattern recognition, sequence modeling, and predictive analytics
Optimization
Mathematical optimization, metaheuristics, and algorithmic efficiency
Complexity Science & Systems
Complex adaptive systems, emergence, and network dynamics
Spatio-Temporal Intelligence
Geographic AI, spatial-temporal modeling, and location-aware systems
Bandits & Online Policy Evaluation
Multi-armed bandits, contextual bandits, and policy evaluation for sequential decision-making
Simulations
Monte Carlo methods, agent-based modeling, and computational simulations
Foundational Mathematics
Core mathematical foundations
Mathematics & Statistics
Linear algebra, multivariable calculus, and statistical inference
Probability Theory
Probability distributions, stochastic processes, and random variables
Information Theory
Entropy, mutual information, and information-theoretic bounds
Advanced Calculus
Optimization theory, variational calculus, and differential equations
🎨
Design & Systems
Human-centered AI design
Systems & Design Thinking
Human-centered design, system architecture, and scalable AI solutions
Psychology & Cognition
Cognitive science, human decision-making, and AI-human interaction
Behavioral Economics
Behavioral biases, decision theory, and economic modeling
Econometrics
Statistical methods for economic data and causal inference
North Star Goals
Ultimate mastery objectives
Agentic Decision Science
How human and business decisions can be nudged and improved with AI and data science
Applied AI/ML Algorithms
Exploring how AI/ML algorithms can be used for specific decision-making problems
Foundational AI/ML Experiments
Conducting experiments around the building blocks and breakthrough concepts in AI/ML
11 Core Topics
4 Foundation Areas
4 Design Principles
3 North Star Goals