About CCS
Two builders. One engagement. Both halves of the work.
TBD — a short positioning paragraph that captures who we are and why we exist. Speak to the reader like a thoughtful prospect, not a marketing audience.
Our Story
How CCS started.
Placeholder
TBD — the origin narrative. How did Rohan and Jatan meet? What was the first problem that became the template for everything that followed? Three or four paragraphs, written like a letter.
Team
The people you'll actually work with.
Rohan Mehta
Chief Engineer
Research Engineer at BlockScience, leading AI integrations into deterministic, expert-system workflows — building reliable ways to ship AI inside production systems. Speaker at EthCC Paris 2023 on LLM integrations into data-science workflows — before most of the category existed — and since at Devcon Thailand and other international conferences. Author of an open-source natural-language interface for cadCAD. Physics and electronics engineering at BITS Pilani; research roots at TIFR on terahertz spectroscopy.
LinkedIn →
Jatan Mehta
Growth & Engagements
Principal Advisor at FundEnable's investment-banking and fundraising practice, after running its Consultancy & Financial Services arm for over three years. Built 50+ VC relationships and prepared investor-ready plans, models, and valuations for 15+ companies; evaluated 500+ startups and mentored 50+ through incubator partnerships. CFA Level III, engineering at DJSCE. Brings financial modeling, engagement scoping, and acquisition research to CCS.
LinkedIn →
Vedant Malpani
BD & Strategy
Corporate lawyer at Trilegal — special situations, structured finance, and debt restructuring. Formerly at AZB & Partners on the Tata Sons – Air India acquisition and cross-border VC work. Brings engagement structuring, compliance automation, and regulated-industry access as a CCS vertical.
LinkedIn →Open Research & Tools
Public work, in the territory adjacent to ours.
A slice of what the team writes and builds in the open — mostly around AI interpretability, retrieval, and the reasoning disciplines that make production AI reliable.
Latent Topologies
Pythongithub.com/rororowyourboat/latent-topologies
Studying LLM latent-space topology using persistent homology and Hodge decomposition. Finds that the Hodge gradient potential recovers the animacy hierarchy — a linguistic universal — without any supervision.
Mechanistic Interpretability
Pythongithub.com/rororowyourboat/mechanistic-interpretability
Head-level analysis of GPT-2-small: tracing how syntax, semantics, factual recall, and pragmatics are actually represented inside the network.
Information Retrieval
Notesgithub.com/rororowyourboat/information_retreival
Information retrieval as epistemic architecture — notes on ontological commitments, bounded attention, and how to assemble context so downstream reasoning stays reliable.
Structured Antagonism
Pythongithub.com/rororowyourboat/structured_antagonism
A reasoning methodology for domains where ambiguity is expensive. Encodes the structural properties of rigorous thinking — design reviews, research plans, architecture decisions — directly into the process.
More at github.com/rororowyourboat.
What We Believe
Three principles that shape every engagement.
Principle 1
TBD — a belief we hold about how the work should be done.
Principle 2
TBD — another conviction that shapes every engagement.
Principle 3
TBD — the thing we refuse to compromise on.