# Prabakaran Chandran — pracha.me # LLM-readable site index for AI assistants and generative engines ## Who Prabakaran Chandran (பிரபாகரன் சந்திரன்) — Tamil Nadu → New York. MS Data Science, Columbia University. Intelligence is becoming abundant; he works on what happens downstream of it: which problems are now worth attempting, and what it takes to build the systems that attempt them. Available December 2026 for Applied Scientist, Researcher & Research Engineer roles. ## The work divides three ways - Choosing — which problems are consequential enough to attempt, and how we would know (/decision-engineering/) - Building — the systems that attempt them, deployed where decisions actually get made (/projects/) - Understanding — the research underneath: probabilistic and causal inference, complex systems, safety (/frontier/) Six and a half years across Mu Sigma, Captain Fresh, Informatica and Nile; research at Columbia and through FAIRE. ## What he is working through 01 The art of problem discovery and solving in the age of superintelligence — immediate and future problems 02 Foundations of frontier intelligence 03 Cognition, philosophy, science and engineering, and the economics of AI 04 Alignment, safety, measurement and control; human–AI interaction; complex systems 05 Data science and machine learning 06 Enterprise problems beyond agent automation 07 Generalists, sovereign AI, and the AGI world 08 Biology, disease, pandemics and biosecurity in the age of AI 09 Global health and wellbeing — where advanced AI can actually move outcomes 10 Longtermism — weighing problems by how far forward their consequences run ## Open Questions - What does a generalist become in the age of AI, and where does that still matter? - Which problems are worth focusing on under superintelligence — and how would we know? - Where do misalignment, malfunction and security threats actually emerge in advanced systems? - What holds epistemic sanity in place when knowledge gets cheap to produce? ## Intellectual Convergence Control science + complexity science + AI → decisions & discovery. Control: state, dynamics, feedback. Complexity: emergence, nonlinearity, adaptation. AI: making both tractable at scale. ## Active Programs - FAIRE (Frontiers in AI Research and Engineering): self-designed, self-directed sprint-based research. Each sprint produces code, a finding, or a documented failure. ## Courses Built autodidactically and in the open at /courses/ — each page grows with readings, notes, and code/projects/assignments. - Research Engineering Alchemy — turn important questions into rigorous research, experiments, systems and evidence. - The Art of Problem Solving in the Age of Superintelligence — identify, formulate and solve consequential problems shaped by increasingly powerful AI. - Complexity Science and AI — understand complex adaptive systems and use AI to model, predict, decide and transform them responsibly. - Open Models, Sovereign AI, Systems, Scale and Product Research and Deployment Engineering — convert learning and projects into public updates, open-source artifacts, reusable tools and community feedback. - AI Safety, Alignment, Human & AI Systems, Evals, Diffusion and Futures — build and evaluate capable AI systems that are safe, aligned, human-centered and understood within broader technological and societal futures. ## Site Sections Navigation is four sections: AI, Problem Solving, Musings, About. - / — the positioning, the open questions, and the latest articles, studies and musings, and the builds - /frontier/ — AI: the research. Deep dives and surveys, a foundations library, the FAIRE programme - /decision-engineering/ — Problem Solving: which problems are worth solving, with decision engineering as the argument (4 pieces) - /musings/ — essays and first-hand experience (55 posts) - /me/ — About: the full narrative, and the way into Courses, Studies, Builds and the Résumé - /studies/ — short, grounded studies from reading and thinking; each begins with a question, why it matters, its source and how sure I am - /primers/ — short introductions to the ideas underneath AI, machine learning, complexity science and the maths around them - /articles/ — in-depth pieces: technical deep dives, project reports, case studies and reviews, listed on AI and Problem Solving - /daybook/archive/ — the earlier Daybook, July to September 2026 - /projects/ — builds (8) - /library/ — working notes on maths, theory, philosophy and control (7) - /courses/ — five courses, each with readings, notes and code/projects/assignments - /curriculum/ — structured learning tracks - /resume/ — career record ## Curriculum Tracks 01 AI — pre-training, SFT, RLHF/RLAIF, instruction following, tool use, agentic architectures, multi-agent systems, evaluation, interpretability, alignment & safety 02 Generative Modeling — autoregressive models, VAEs, GANs, normalizing flows, diffusion, score matching, flow matching, optimal transport, energy-based models 03 Representation Learning — self-supervised pretraining, contrastive learning, masked autoencoding, JEPA, disentangled representations, multimodal alignment, vision-language models 04 Neural Networks & Deep Learning — CNNs, RNNs, transformers, GNNs, optimization, training dynamics, normalization, scaling laws, NAS, theory 05 Statistical & Probabilistic ML — Bayesian inference, MCMC, variational inference, graphical models, Gaussian processes, EM, uncertainty quantification 06 Reinforcement Learning — MDPs, dynamic programming, Q-learning, policy gradients, actor-critic, PPO, model-based RL, world models, offline RL, RLHF, multi-agent RL 07 Attention, Memory, Reasoning and Continual Learning — attention mechanisms, transformers, long-context, chain-of-thought, RAG, Hopfield/NTM/DNC, continual & lifelong learning, catastrophic forgetting 08 Causal and Statistical Inference & Modeling — SCMs, potential outcomes, treatment effects, IV, DiD, regression discontinuity, causal discovery, counterfactuals, causal representation learning 09 Algorithms and Systems for AI — GPU programming, distributed training, quantization, pruning, distillation, efficient architectures, inference serving, ML compilers (XLA, Triton), MLOps 10 Complexity, Cognition and Natural Intelligence — complex dynamical systems, nonlinear dynamics, emergence, adaptive control, cognitive science, neuroscience, embodied cognition, scientific discovery, life/biological/physical sciences ## Contact & Profiles - Site: https://pracha.me - GitHub: https://github.com/prabakaranc98 - LinkedIn: https://www.linkedin.com/in/prabakaranchandrantheds/ - Twitter/X: https://x.com/PraCha98 - Email: pc3197@columbia.edu ## Feed & Indexing - RSS: https://pracha.me/feed.xml - Sitemap: https://pracha.me/sitemap.xml - Robots: https://pracha.me/robots.txt (AI crawlers explicitly welcomed)