Harmonic
Website:
harmonic.ai
Job details:
VDIP & Lextron (will be housed under a New OpCo to be incorporated) is building India’s first structured AI corpus for venture deal intelligence and the legal infrastructure platform built on top of it. We are pre-seed, founder-led, and closing our first institutional round in Q3 2026.
Two products, one compounding data engine. Both are in active development. The full product brief — architecture, market, and technical approach — is shared under NDA after the first conversation.
ROLE AT A GLANCE
Title
Co-Founder & CTO
Entity
[To be incorporated]
Location
New Delhi (preferred) or Remote India — with availability for in-person meetings
Equity
Significant founder-level stake. 4-year vesting, 1-year cliff. Specifics negotiated based on experience and timing.
Cash compensation
Below-market through pre-seed. Market-rate from seed close. This role is an equity play, not a salary play.
Reporting to
Vatsal Gaur, Founder & CEO
Hiring stage
Pre-seed. You are the second co-founder, not the first technical hire.
THE PROBLEM SPACE
India’s venture ecosystem generates thousands of structured legal transactions annually — seed instruments, shareholder agreements, term sheets, Series A and B documents. That transaction data is almost entirely unstructured and inaccessible at scale. The intelligence locked inside it — deal economics, market benchmarks, clause-level patterns — has never been extracted, structured, or made queryable.
The CTO will own the technical architecture that changes that. The core challenge is a domain-specific AI problem: structured extraction from specialist Indian legal documents, at scale, with the precision required for a data product that institutional clients will depend on.
The second product — Lextron — is the infrastructure platform built on top of that intelligence layer. AI-generated, advocate-reviewed legal documents for founders and companies who cannot access institutional legal services at current price points.
What we are not building: A general-purpose legal AI tool. A chatbot for contracts. An LLM wrapper. We are building a proprietary data product with a domain-specific extraction layer on top. The corpus is the moat. The AI is the delivery mechanism.
WHAT WE NEED (NON-NEGOTIABLE)
Technical depth:
– Strong NLP/ML background with hands-on experience building domain-specific extraction pipelines — not just fine-tuning general models, but designing structured extraction from specialist document types.
– Production-grade Python and backend engineering. You have shipped real products, not prototypes.
– Experience with LLM application architecture — RAG pipelines, vector databases, embedding strategies, prompt engineering at scale. You know when to use a large model and when not to.
– Cloud infrastructure — AWS preferred, data pipelines, API design. You make decisions that hold at 10x scale.
Founder mindset:
– You have founded a company, been an early technical employee at a startup where you owned the full stack, or led a product at a company with direct customer exposure. You know what it is to build with no safety net.
– You are comfortable with ambiguity and architectural change. Pre-seed means the product will evolve. You design for adaptability.
– You can explain the technical architecture to non-technical investors and institutional clients. The CTO will be in the room for partnership conversations, not just engineering standups.
WHAT GIVES YOU AN EDGE
– Prior experience in legaltech, fintech, or any domain requiring structured extraction from unstructured professional documents.
– Some familiarity with Indian corporate or VC legal instruments. You don’t need a law degree, but you should be able to read a term sheet without asking what a CCD is.
– Experience building data products where the corpus or proprietary dataset is the core moat — you understand the difference between a data company and a model company.
– Prior work with Indian data protection or compliance frameworks in a technical capacity.
– Previous co-founder or founding engineer experience at a company that raised institutional capital.
THE EQUITY CONVERSATION
Why this matters: The technical co-founder is not a service provider to this product. The intelligence value of what we are building only exists if the right person builds the extraction layer. Equity reflects that, not the other way around.
Equity stake will be meaningful and founder-level. The exact percentage reflects when you join, the depth of your technical contribution, and your long-term commitment. We are not offering a modest ESOP. We are looking for a co-founder who will look back in 10 years and say they built this.
Standard terms: 4-year vesting, 1-year cliff, acceleration on change of control. Anti-dilution protection on future raises.
THE RIGHT PERSON
We are not running a standard hiring process. We are looking for one person. That person probably looks like this:
– 5–12 years building technical products. NLP/ML experience in the recent years is important; the rest can be varied.
– Has worked in India or deeply understands the India tech and startup ecosystem.
– Is at an inflection point — leaving a corporate role, wrapping up a previous startup, or actively looking for the right co-founder opportunity. Not passively interested.
– Believes that in AI products, the moat is the data and the domain — not the model. Is looking to build something where that conviction is the foundation.
– Can start within 60 days of agreeing terms.
HOW TO APPLY
This is a co-founder search, not a job application. The process is a conversation, not a filter.
– Reach out to Vatsal Gaur directly: vatsal@reblock.world with subject line “CTO Co-Founder — [your name].”
– Include: what you have built (links, not CVs), why this problem specifically, and one concrete technical opinion on what you think is the hardest part of building a domain-specific NLP extraction layer for specialist legal documents.
– First conversation: 45 minutes. No deck, no formal pitch. A technical and strategic conversation about the problem and whether you are the right person to solve it.
– Full product brief, technical architecture, and financial context shared under NDA before the second conversation.
– If both sides have conviction: co-founder agreement and equity conversation.
One thing: We will wait for the right person. If you read this and thought ‘this is the problem I want to work on’ — reach out. If you are not sure, it is probably not right.
Tips: Provide a summary of the role, what success in the position looks like, and how this role fits into the organization overall.
Responsibilities
[Be specific when describing each of the responsibilities. Use gender-neutral, inclusive language.]
Example: Determine and develop user requirements for systems in production, to ensure maximum usability
Qualifications
[Some qualifications you may want to include are Skills, Education, Experience, or Certifications.]
Example: Excellent verbal and written communication skills
Click on Apply to know more.