Kiriti & Aish

Kiriti & Aish

38K Subscribers

AI Founder & Advisor to F500s | Ex-AWS

Maven's Terms and Privacy Policy.

Contact

Kiriti & Aish will make you a problem-first AI builder

Aishwarya Naresh Reganti is the founder and CEO of LevelUp Labs, a human-first AI startup that helps develop and deploy production-ready generative AI solutions for mid to large enterprise clients. With over 9 years of experience in machine learning, she has published more than 35 research papers at top-tier AI conferences, including NeurIPS, AAAI, and CVPR.

Kiriti Badam is a member of the technical staff at OpenAI Codex, with over a decade of experience designing high-impact enterprise AI systems. He specializes in AI-centric infrastructure, with deep expertise in large-scale compute, data engineering, and storage systems.

Worked/Taught At
MIT
University of Oxford
Amazon Web Services
Microsoft

Alumni reviews

Beyond Evals: Designing Improvement Flywheels for AI Products

The course was a perfect blend of theory and hands on exercises. Everything from the skills, agents and datasets curated and provided during each of the weeks were very helpful in learning through practice. I also appreciate the focus on understanding where human judgment is needed and where AI can help augment.
Reviewer profile image

Aiza

June Cohort
AI Product Manager · Cisco

Beyond Evals: Designing Improvement Flywheels for AI Products

Aish and Kiriti have done it again and nailed a topic that i have been iterating it for a while. They made it simple and equipped you with latest tools and scripts that you can use it from Day 1 from joining this course. I already used it for 2 of my projects and now revisiting the course content again..

Puja

June Cohort
Sr. Product Manager · undisclosed

Beyond Evals: Designing Improvement Flywheels for AI Products

Exceptionally well-structured and perfectly paced. Three weeks is exactly the right length for a practitioner-level deep dive, and the curriculum covers serious ground: offline and online evals, multiple architectures, and — what I appreciated most — real coverage of agent harnesses and skill-based automation. This course is running alongside the field, not behind it. Aish's presentations are dense with useful frameworks and current research, the kind of material you keep referencing long after. The combination of theory and hands-on assignments added genuine depth that passive learning can't replicate. What sets this apart is Aish and Kiriti themselves — their presence and responsiveness in the live sessions makes questions spark real discussion rather than disappear into a void. I came away with a much stronger foundation for building and thinking about evals. Already looking forward to their next course.
Reviewer profile image

Trina

June Cohort
Founder & Product Manager · TakeControl

Beyond Evals: Designing Improvement Flywheels for AI Products

Great Course!! What I liked about this course is the way it goes beyond "LLM-as-a-Judge" and gets into Reference data sets, Architecture specific evaluation and evaluation in Production.

Balaji

June Cohort
Data & AI Engineering Business Unit Leader · Sigmoid

Beyond Evals: Designing Improvement Flywheels for AI Products

Most teams know how to run evals. Very few know how to turn eval signals into a real improvement loop. Aishwarya and Kiriti's Beyond Evals course is the best treatment I've seen of that missing link — offline evals, online evals, and production monitoring wired into one data flywheel. As someone building eval-gated agentic systems in production retail, I found it directly applicable from week one. Highly recommended for anyone serious about shipping AI products that get better over time. What sets 'Beyond Evals course' apart is that it treats evaluation as a system, not a checklist. The course connects offline evals, online evals, and production monitoring into a single loop where production signals feed back into your datasets and eval corpus. That framing changed how I think about my own architecture: evals aren't a gate you pass once — they're the mechanism that earns your system more autonomy over time, and the data flywheel they generate is the real moat.
Reviewer profile image

Biplab

June Cohort
Principal · Infusion Advisory Solutions