About the job
About Fermi
Fermi is an innovative AI tutoring startup, launched by Meraki Labs, committed to transforming education on a global scale. Our vision is to democratize access to exceptional tutoring, empowering every student through state-of-the-art AI technology. We are a passionate and agile team of educators, designers, and engineers, driven by the belief that intelligent design and advanced AI can significantly enhance the learning journey.
Under the leadership of seasoned founders, including Mukesh Bansal (Founder of Myntra, CureFit, and Nurix) and Peeyush Ranjan (VP of Google, former CTO of Flipkart and Airbnb), we are developing a product aimed at reshaping the educational landscape for millions of students, transforming academic obstacles into achievements. Operating from Bangalore, India, we cultivate a fast-paced environment that emphasizes impact, ownership, and direct collaboration to shape the future of AI in education.
Role Overview
We are seeking a dedicated AI Engineer who excels as a problem-solver: someone who can develop, debug, and iterate rapidly. You’ll play a key role in constructing the foundational AI workflows that drive our tutoring product, particularly focusing on agentic tutoring systems, and ensuring they are robust, scalable, and production-ready.
This position is designed for those who thrive in dynamic environments. If you prefer tightly defined tasks or roles limited to either research or backend work, this may not be the right fit. However, if you enjoy building comprehensive systems from the ground up and witnessing your contributions impact production swiftly, this is the role for you.
What You’ll Do
Develop and deploy core AI workflows for our tutoring application:
Agentic tutoring flows (hinting, step-by-step guidance, misconception detection)
Answer evaluation, grading, and feedback mechanisms
Content retrieval and grounding (ingestion, chunking, embedding, re-ranking)
Personalization features (student memory, progress tracking, difficulty adjustment)
Multimodal processing pipelines (including images/diagrams; voice experience is a plus)
Transform prototypes into stable production systems:
Optimize for latency and costs (caching, batching, streaming, fallbacks)
Iterate based on evaluations (offline tests, regression checks, quality assessments)

