Logic Design Engineer

Company Description
Moleculyst Ventures Private Limited is a deep-tech AI company building a reasoning checker for natural language. Our goal is to make LLM output checkable, so that people only review the cases that need them. We break text into small steps, verify each step deterministically against source documents such as contracts and laws, and send only the uncertain cases to a human. The work combines formal logic, computational linguistics, and machine learning in a small research-driven team.
Role Description
The Logic Design Engineer will design and build the logical core of our checker: the proof system that decides whether a step written in English follows from its premises. Day-to-day work includes:
• Defining inference rules and axioms over first-order logic
• Specifying how English function words (and, or, if, every, not) map to logical operators
• Proving or testing that the rule set is sound
• Implementing the step-by-step proof checker and testing it against established solvers
• Working with our linguists on the mapping from sentence meaning representations to logic
This is a full-time, hybrid role based in Delhi, India, with part of the work done from home.
Qualifications
• Strong foundation in mathematical logic: first-order logic, proof theory, soundness and completeness
• Working knowledge of proof systems (Hilbert-style, natural deduction, sequent calculus, or tableaux)
• Experience with automated reasoning tools such as SAT/SMT solvers or theorem provers
• Good programming skills, with the ability to implement a checker from a formal specification
• Bachelor's or Master's degree in Computer Science, Mathematics, Logic, or a closely related field
• Familiarity with formal semantics of natural language (discourse representation theory, Montague-style semantics) is a plus
• Experience with proof assistants (Lean, Coq, Isabelle) or controlled natural languages is a plus
• Clear written reasoning and attention to detail; able to work closely with linguists and ML engineers

Skills
Machine Learning, LLM

Work arrangement
Hybrid