Apply for Gen AI Researcher – Reasoning & Benchmarks

Role: Gen AI Researcher – Reasoning & Benchmarks

Function: AI Research

Location: Bangalore

Type: Full-time

Industry: Information Technology & Services

About Company

A research-first AI company incubated at the Indian Institute of Science (IISc). The company builds advanced AI for the planning and operations of critical networks.

Its core technology is a World Model that augments LLMs to reason about network state and dependencies, even with limited or noisy data. The founding team has raised over $100M in prior ventures and published at leading AI research venues.

Having just closed its seed round, the company is assembling a team of ambitious, frontier-focused researchers.

Position Overview

This role sits at the core of the company's research agenda — designing and building reasoning benchmarks across science and engineering domains to rigorously evaluate AI systems at the frontier of reasoning and planning. You will build evaluation infrastructure, develop agents, and iterate on them against real-world science and engineering tasks. You'll work directly with the founders and IISc faculty in a lean, research-driven environment where your output shapes the company's technical direction.

Role & Responsibilities

• Design, build, and validate reasoning benchmarks spanning science and engineering domains, ensuring rigor and reproducibility
• Develop evaluation pipelines using eval harnesses and benchmark suites (e.g., SWE-bench) to systematically assess AI system performance
• Build and iterate on agents capable of tackling complex reasoning and planning tasks, guided by benchmark results
• Contribute to the company's core research agenda — document methodologies, findings, and experimental results to publication-ready standards
• Collaborate closely with IISc faculty, and internal research, agent, and data teams to align benchmark design with real-world network planning problems
• Engage with the open-source research community to stay current with and contribute to frontier developments in reasoning and evaluation

Must Have Criteria

• BTech from a top-tier institute (IITs/IISc/BITS or equivalent) with 2+ years of relevant AI/ML research experience, OR MTech/PhD from a top-tier institute
• Solid foundations in probability, statistics, and machine learning — demonstrable through coursework, research, or published work
• Hands-on experience with evaluation harnesses and benchmark suites (e.g., SWE-bench or comparable frameworks)
• Strong coding and software engineering skills sufficient to build and maintain reproducible research pipelines
• Demonstrated ability to build and iterate on AI agents for reasoning or planning tasks

Nice to Have

• Prior publications at leading AI research venues (NeurIPS, ICML, ICLR, ACL, or equivalent)
• Prior experience designing or contributing to benchmark datasets or evaluation frameworks
• Familiarity with network planning, operations research, or science/engineering simulation domains
• Experience working in early-stage or research-lab environments with high autonomy

What We Offer

• Direct collaboration with IISc faculty and a founding team with a track record of $100M+ raised and frontier AI publications
• Ownership over a core research function at a seed-stage company building at the frontier of reasoning AI
• Opportunity to publish and contribute to the broader AI research community as part of your role
• Lean, research-first culture with high autonomy and cross-disciplinary exposure across agents, data, and network AI