AI Engineer (0 2 years)

📋 Description

• • Join Fermi AI, an early-stage AI tutoring startup born from Meraki Labs, and play a pivotal role in revolutionizing global education. Our ambitious mission is to democratize access to world-class tutoring, making it available to every student through the power of advanced Artificial Intelligence. As a lean and dynamic team comprising passionate educators, skilled designers, and innovative engineers, we are united by a profound belief that thoughtful design, coupled with cutting-edge AI, can fundamentally transform the learning experience for millions.
• • Led by accomplished founders, including Mukesh Bansal (Founder of Myntra, CureFit, and Nurix) and Peeyush Ranjan (formerly VP at Google, CTO at Flipkart and Airbnb), we are developing a groundbreaking product poised to reshape how students learn, transforming academic challenges into significant accomplishments. Our operations are based in Bangalore, India, where we cultivate a high-velocity environment that places a premium on impact, ownership, and direct collaboration, all aimed at defining and building the future of AI in education.
• • We are seeking an AI Engineer who embodies a problem-solver's mindset – an individual capable of shipping, debugging, and iterating rapidly. You will be instrumental in constructing the core AI workflows that power our tutoring product, with a particular focus on agentic tutoring systems. Your responsibilities will extend to hardening these systems into production-grade, scalable, and observable solutions. This role is intentionally designed for individuals who thrive in dynamic, end-to-end development environments and are eager to see their work deployed and making an impact quickly. It is not suited for those who prefer a tightly defined scope, predictable tasks, or exclusively research-focused or backend-only responsibilities.
• • **Key Responsibilities and Contributions:**
• • **Develop and Ship Core AI Workflows:** You will be at the forefront of building and deploying the essential AI functionalities for our tutoring application. This includes:
• • **Agentic Tutoring Flows:** Designing and implementing systems that provide intelligent, context-aware guidance, such as hinting, step-by-step instruction, and sophisticated misconception detection.
• • **Answer Evaluation:** Creating robust mechanisms for grading student responses and providing constructive feedback loops.
• • **Retrieval and Grounding:** Developing pipelines for content ingestion, efficient chunking, embedding, and re-ranking of educational materials to ensure accurate and relevant information retrieval.
• • **Personalization:** Building systems that adapt to individual student needs, incorporating student memory, progress signals, and dynamic difficulty adjustments.
• • **Multimodal Pipelines:** Contributing to the development of pipelines that can process and integrate various data types, including images and diagrams. Experience with voice processing is a significant bonus.
• • **Transition Prototypes to Production-Ready Systems:** Your role will involve taking experimental AI models and workflows and transforming them into reliable, efficient, and scalable production systems. This entails:
• • **Latency and Cost Optimization:** Implementing strategies such as caching, batching, streaming, and fallback mechanisms to ensure optimal performance and resource utilization.
• • **Eval-Driven Iteration:** Employing rigorous evaluation methodologies, including offline test sets, regression checks, and quality gates, to drive continuous improvement and ensure system robustness.
• • **Observability:** Establishing comprehensive monitoring solutions, including traces, logs, metrics, and prompt version tracking, to ensure system health and facilitate rapid debugging.
• • **Reliability and Safety:** Implementing guardrails, ensuring appropriate refusal behaviors, and guaranteeing age-appropriate and policy-compliant output to maintain a safe and effective learning environment.
• • **End-to-End Feature Ownership:** You will take full ownership of features, guiding them from initial rough Product Requirements Documents (PRDs) through implementation, deployment, monitoring, and iterative refinement.
• • **Cross-Functional Collaboration:** Work closely with product design and education specialists to translate pedagogical goals and learning objectives into effective AI behaviors and functionalities.
• • This role is ideal for individuals who default to ownership, possess the ability to move quickly without sacrificing quality, thrive in ambiguous environments by driving clear execution, and view quality as an inherent engineering problem solvable through tests, evaluations, instrumentation, and iteration.