Research Scientist, Applied AI

This is a fully remote position, open to applicants in India.

📋 Description

• Take charge of and enhance the agent harness: the primary loop, tools, environment, guardrails, and model behavior that drive the agent.

• Enhance agent tool selection, context construction, sub-agent coordination, and the management of lengthy, multi-step tasks through methodical experimentation.

• Facilitate model selection across various tasks while balancing the quality of answers, latency, and cost.

• Design and develop evaluations for intricate revenue tasks, including benchmarks, regression detection, and automated evaluation pipelines.

• Create evaluation datasets and grading methods for tasks lacking clear ground truth.

• Develop scoring and prediction models, such as deal health scores, cross-deal pattern mining, and explainable risk signals.

• Lead fine-tuning efforts as the platform transitions into advanced model training.

• Investigate and design reinforcement learning strategies that enhance agent behavior.

• Enhance context engineering and data quality through entity resolution, information extraction, and knowledge and context graphs.

• Manage ambiguous research challenges from start to finish: formulate hypotheses, design experiments, build training, evaluation, and grading infrastructure, and deliver results to production.

• Collaborate closely with the US-based Head of AI and the applied AI team in India.

⛳️ Requirements

• Over 5 years of experience in applied AI research or applied science at a senior level.

• MS or PhD in Computer Science, Mathematics, Statistics, or another quantitative field.

• Practical experience in building and enhancing agentic systems and agent harnesses, including tool usage, multi-step reasoning, and context management.

• Hands-on experience with LLMs or large-scale deep learning systems, including training, fine-tuning, and RLHF.

• Strong understanding of model behavior, focusing not just on usage but also on underlying principles.

• Experience in designing evaluation datasets or frameworks, particularly for tasks without clear ground truth.

• Demonstrated ability to navigate and drive ambiguous research with minimal supervision.

• Ability to work comfortably with a US-based leadership team across different time zones.

• Publications in top-tier peer-reviewed conferences or journals are preferred.

🏝️ Benefits

• Competitive salary and equity at Series A, ahead of the next funding round.

• Remote work opportunity.

• In-person team meetings in Bangalore.

• Meaningful responsibilities from day one: ownership of core agent capabilities.

• Direct collaboration with the Head of AI and a close feedback loop with founders and product teams.

• Opportunity to influence an AI-first product during a critical phase.