Jobs at Pramana Labs

Own end-to-end product design across customer, developer, internal, and shared platform experiences. Translate complex AI verification systems, rules, evidence, and dependencies into understandable workflows. Conduct research, prototype, deliver detailed designs, build design-system foundations, and partner closely with engineering through implementation. Help shape product direction, interaction patterns, accessibility standards, and a scalable visual language in an early-stage AI company.

Conduct model training experiments, analyze results, develop auto-formalization and prover-system harnesses, build evaluation benchmarks, and investigate model successes and failures. The intern will collaborate with researchers and engineers to translate findings into improved experiments and approaches for symbolic reasoning and formal AI systems.

Test a rules-to-proof AI product across rule inputs, formal specifications, prover outputs, proof artifacts, APIs, and proof viewers. Identify functional defects, edge cases, security risks, malformed inputs, conflicting rules, and data-handling issues. Create reproducible bug reports, verify fixes, and progressively automate testing using Python, pytest or similar frameworks, and CI/CD tools.

Build full-stack infrastructure, DevOps, and MLOps tools for AI development teams. Responsibilities include CI/CD, reproducible environments, experiment and evaluation workflows, self-service applications, observability, cost tracking, deployment automation, identity and access services, LLM gateways, and usage metering. Collaborate directly with engineers and researchers to improve internal workflows, reliability, and operational efficiency.

Own a product area from customer discovery through launch and adoption. Define product direction, roadmaps, requirements, user workflows, experiments, and success measures. Translate AI formalization and verification capabilities into understandable, trustworthy products while partnering with engineering, research, design, domain experts, and commercial teams. Measure activation, usage, task completion, time to value, and product quality, and incorporate customer feedback into prioritization.

Conduct research on reinforcement learning post-training for reasoning models. Develop RLVR methods using Lean 4’s verified rewards, create novel RL algorithms and test-time compute optimizations, advance autoformalization, curate training data, and build evaluations for formal reasoning. The role owns the full research loop from rollout sampling and reward design through model updates and evaluation, with emphasis on autonomous, state-of-the-art research in long-horizon logical tasks.

Build full-stack product experiences spanning responsive interfaces, backend services, APIs, data models, integrations, and testing. Integrate AI models and verification services into reliable customer products, collaborate with research, product, and design teams, learn from users, and own software quality, monitoring, security, performance, and production support.