Maxim

It’s always great to chat with Rohit Pandharkar. He and his team is doing great work at EY India on enabling GenAI in different verticals. Maxim is proud to be the AI Quality partners for his team to ship faster and better ??

Over the past year, we've witnessed AI become a mission-critical priority for organizations around the world. From RAG-based QA systems to multi-agent architectures, adoption continues to soar. However, one consistent challenge echoes across all AI development efforts: measuring and improving the quality of these complex AI systems. That’s why we've been hard at work building Maxim—to make it incredibly simple for modern AI teams to evaluate and enhance their AI application quality, from experimentation → pre-release testing → production. Our team is committed to delivering the slickest developer experience for AI evaluation and observability. With that, we're live on Product Hunt with our self-serve AI evaluation offering ???? I invite you to try out Maxim and share your feedback with us, so we can continue enhancing your experience as you push the boundaries of AI applications. Check us out here: https://lnkd.in/gQYar3ah. We’d love your feedback and support ???? Sign-up for a free trial: https://lnkd.in/gg6nrkD2 Book a demo with us: https://lnkd.in/g3bzTmxV Happy building ????????

During my role at Postman, I had the good fortune of exploring several LLMOps platforms - covering observability, evaluations, and more. While many tools focus on specific aspects, none seem to deliver everything that’s essential for the complete AI development lifecycle (AI DLC). That’s why I truly "admire" the team at Maxim for nailing the most critical pieces. Here’s what stands out to me: 1?? End-to-end lifecycle coverage: From pre-production experimentation to in-production monitoring and post-production iteration via a complete evaluation and data flywheel, Maxim has built a platform that truly supports every stage of AI development. 2?? Beyond granular LLM-call level eval/logging: Maxim works easily with your custom endpoints and services, enabling support for even complex multi-step AI workflows. This is so critical, however most platforms miss it? 3?? A complete evaluation and data flywheel: They’ve absolutely nailed it. Whether it’s: - Custom evaluators with native runtime execution/BYO eval API, - Manual human evaluation via UI dashboards, - Real-time eval triggers, or - Pre-production experimentation, they’ve covered all bases for AI product teams. 4?? A low-code, canvas-based playground: Perfect for designing both traditional DAG workflows and trending Agentic workflows - any multi-step data/control flow really. This is an exceptional example of Problem-Solution Fit and User-Centered Design. Great work folks - Akshay Deo Vaibhavi Gangwar (VG) Sapna Nayak and others!

We’re live on Product Hunt with our self-serve AI evaluation platform ????Ship high-quality AI applications, much faster, with Maxim! Maxim’s end-to-end evaluation and observability platform empowers teams to test and improve their AI products across the development lifecycle: experimentation → pre-release testing → production. With Maxim, you can: ? Iterate fast with LLMs and experiment across your prompts, data, tools, architectures ?Test your agentic AI systems easily with a simple API endpoint ? Monitor AI quality in production and establish the critical feedback loop from production → iteration ?Curate high-quality datasets across the development lifecycle …and lots more Check us out here: https://lnkd.in/gDw_BAmq . We would appreciate your feedback and love on Product Hunt ?? Happy building ??

?? Only 3 DAYS to go! On the 21st of November, Maxim AI will be going live on Product Hunt and we would really love to have your support! ?? Our team at Maxim is committed to building the best developer experience for AI evaluation and observability. The platform helps you streamline your AI development lifecycle—from experimentation to post-release monitoring. Stay in the loop by clicking "Notify Me" on our Product Hunt page (link in comments). ?Your support will play a key role in helping us deliver even more value to the AI community. P.S If any of you are building with GenAI, would be happy to demo or chat about your use cases separately as well.. do reach out :)

One common theme we hear from our users is: effectively balancing auto-evaluations with human feedback to drive both the quality of their AI workflows and the efficiency of their development processes. At Maxim AI, from day one, we have built our platform anchored on the vital feedback loop from human reviews to product iteration, helping AI teams align their AI experiences with human preference. Teams can either self-annotate evaluation runs and logs or set sampling logic to send to subject matter experts (SMEs) for evaluations. The end-to-end workflows for human evaluation, request → feedback collection → feedback analysis and utilization, have been a game-changer for teams evaluating their AI applications on Maxim. A few hero use cases teams use these workflows for: ? Create custom evaluation criteria and request feedback based on any sampling logic from SMEs via our dedicated Reviewer Portal. ? Track and follow up on the status of human reviews from your experiments page with a single click. ? Curate golden datasets based on human-corrected outputs by SMEs. These golden datasets can be used to improve your evaluation or fine-tuning workflows, or can simply be used for in-context learning. ? Create human review queues from production logs and assign them to teammates for annotation. ...and lots more. If you’re curious to learn more about how Maxim is helping teams to the last mile of productionizing AI applications, I would love to chat with you. Book a demo: https://lnkd.in/g3bzTmxV Request access for a free trial: https://lnkd.in/gGzvmxA5

Exciting news ?? ! Manav Singhal from our team is presenting his work on NoFunEval, a benchmark evaluating code LMs on non-functional requirements, at COLM24 conference. Connect with Manav Singhal to discuss evaluations and code LMs! https://lnkd.in/g9r5WGst

Maxim AI's Aniruddha Chattopadhyay and Parth R. will showcase how our evaluation and observability platform empowers modern AI teams to ship applications with quality, reliability, and speed at the MagicBall AI event hosted by Grayscale Ventures. At Maxim, AI quality is at the core of our mission, and we will demonstrate how our platform ensures application quality throughout the development lifecycle. Join us at the event on September 30th, 2024, to experience our technology firsthand. We look forward to connecting with fellow GenAI builders to discuss emerging development workflows and challenges. ?? Catch us live at the event to see our tech in action. We look forward to connecting with fellow GenAI builders to discuss emerging development/evaluation workflows and challenges. #OnlineEvaluations #MaximAI #DevTools #GenAI #LLMs

As Maxim AI's user base grew, we noticed how vital spreadsheets are in ML engineering workflows. So we rolled up our sleeves and built a Dataset feature that's not just compatible but seamlessly integrates with spreadsheets. And this is not a marketing gimmick! See the video :) Here's what makes it special: ? Uses canvas APIs - absolute zero DOM manipulations ? Handles millions of rows easily ? Efficient data exchange And the biggest of all - ITS MULTIMODAL!! We have put a lot of thought into this to ensure it helps ML/AI engineers in their daily work. It's been rewarding to see it come together. If you are building using #LLMs - Try our 14-day free trial here: https://lnkd.in/dRGrkHyz.?#GenAI #Agents #Evals

Manav Singhal and Aniruddha Chattopadhyay are cooking up something cool for Maxim AI. We're tackling the complexity of agentic workflows head-on, focusing on streamlining build and evaluation processes. Our goal? Making continuous evaluation a breeze. We can't wait to ship this in the coming months; it will be a game-changer! If you are building agentic workflows, Sapna Nayak , Aniruddha Chattopadhyay, and Manav Singhal are diving deep into these challenges and would love to learn about your experiences. If you are building using #LLMs - Try our 14-day free trial here: https://lnkd.in/dRGrkHyz. #GenAI #Agents #Evals