Tata Sons vs Tata Trusts: Ownership, Control & the IPO Fight
Why the RBI wants Tata Sons listed under Scale-Based Regulation, why Tata Trusts resists, the holding company discount dilemma, and the structural concentration risk in TCS.
Writing & Reflections
Notes from the build.
A growing library of notes on AI engineering, financial systems, product thinking, research, and the ideas I'm exploring along the way.
Featured Analysis
5 min read
Tata Sons vs Tata Trusts: Ownership, Control & the IPO Fight
Why the RBI wants Tata Sons listed under Scale-Based Regulation, why Tata Trusts resists, the holding company discount dilemma, and the structural concentration risk in TCS.
Selected Notes
9 notes published
What NSFR Actually Tells You About Bank Structural Liquidity
Why the Net Stable Funding Ratio matters more than short-term liquidity coverage when evaluating structural balance-sheet stability and funding mismatches.
Why Market Validation Trumps Vibes: A Three-Step Query Framework
Stop writing production software before knowing whether active search demand exists. A practical framework to validate buyer intent before committing capital.
Distilled Video Models & Regional AI Production Economics
How specialized, distilled generative models bring video synthesis down to a fraction of a cent per second, revolutionizing regional commerce.
Physical AI & Heavy Industry: Where Software Speed Hits Hardware Bottlenecks
Training foundation models on real-world physics and molecular dynamics enables automated engineering, but software iteration speed inevitably confronts physical validation cycles.
The Shift from Static Machinery to Robot Fleet Configuration
How reconfigurable robotics and localized micro-factories are transforming manufacturing roles from manual assembly to intelligent fleet orchestration.
Public Market Listings as the Ultimate Stress Test for AI Valuations
Private venture rounds can sustain narrative-driven multiples, but public equity markets enforce rigorous discipline on gross margins and recurring cash flow.
Designing Deterministic Validation Schemas for Multi-LLM Pipelines
Lessons from production AI pipelines: why multi-LLM orchestration fails without rigorous schema guardrails and programmatic retry backoff.
Graph Neural Networks for Topological Food-Web Interactions
Research notes from OIST: applying GraphSAGE and GAT architectures to model topological trophic networks and synthesize scientific literature.