Introducing the World’s First AI System for ML Engineering 

Beats OpenAI’s O1 in their very own MLE benchmark with a rating of 26% vs. 16.9%

NEO automates machine studying engineering from begin to end, lowering experimentation, coding and operational grunts by 1000’s of hours

PALO ALTO, CA / ACCESS WIRE / November 21, 2024 / A stealth startup publicizes NEOThe world’s first autonomous AI engineer to automate your complete machine studying workflow for ML engineering, giving ML builders superhuman expertise to construct AI fashions in document time and save 1000’s of hours of grunt labor. Some giant, multi-billion-dollar enterprises are working with NEO as early design companions. A waiting list is open to early adopters.

Powered by a proprietary multi-step reasoning algorithm that gives end-to-end automation, NEO permits firms to have a software that actually competes with the best-in-class ML engineer. The system permits customers to talk in pure language, performs knowledge engineering and may run a whole bunch of refinement experiments. Demo video is out there here.

Meet NEO – First Autonomous Machine Studying Engineer

In a serious breakthrough, the founding crew examined NEO on the MLE benchmark, an OpenAI benchmark to measure how effectively AI brokers carry out in machine studying engineering duties. NEO scored 26% with smaller fashions akin to GPT 4o and Sonnet 3.5, in comparison with OpenAIs 16.9% utilizing the bigger and costlier o1 mannequin with the AIDE scaffolding, bringing the NEO AI engineer to the extent of a Kaggle Grandmaster.

Constructing high-quality AI fashions and pipelines requires complicated ML engineering and vital experimentation. Machine studying duties embrace cleansing and merging knowledge, refining fashions with completely different mixtures of hyperparameters and mannequin sorts, and performing evaluations earlier than ultimate deployment. An optimized pipeline typically contains quite a lot of frameworks, together with vLLM, TensorRT, and Hugging Face transformers.

“Most present ML developer instruments remedy solely a part of the worth chain, akin to coaching, refinement, or deployment,” mentioned Saurabh Vij, CEO and co-founder. “Think about having an ML engineer who works tirelessly, by no means takes a break, prices a fraction of the value, and handles all of the repetitive experimentation and work on the infrastructure degree. That is what we’re constructing – the primary autonomous ML engineer – a game-changer for firms struggling to scale their AI efforts.”

“AI automations are indispensable instruments for all builders and companies seeking to get essentially the most worth for his or her money and time,” Saurabh added. “It is one thing that machine studying would not have and desperately wants as we speak. ML engineers are very scarce and costly. Moreover, the various components of experimentation, retraining and redeploying suck the soul out of ML groups and decelerate your complete innovation cycle.”

NEO is a multi-agent system that completes duties in a number of phases:

• Section 1: The NEO understands the issue and explores a number of concepts to resolve it.

• Section 2: Utilizing its multi-step reasoning capacity, NEO first examines all potential options after which creates the plan for essentially the most promising path and conducts rigorous experiments for various steps. These embrace:

  • Remodel datasets

  • Nice tuning

  • Rag

  • Stake

  • Framework integrations

• Section 3: NEO screens the standard of the mannequin when it comes to latency and accuracy, in addition to the well being of the infrastructure when it comes to node load balancing and prices.

Co-founder Gaurav Vij mentioned, “We based NEO to tackle the duties that distract the artistic builders from innovating and losing a lot time. In consequence, this isn’t an answer to switch folks; they’re at all times busy guiding him and monitoring his habits. Similar to a brand new worker, you’ll be able to prepare NEO and he’ll get to know your processes, get to know the script and act accordingly.”

“Immediately there are fewer than 600 Kaggle Grandmasters” continued Saurabh. Our long-term imaginative and prescient is to convey an AI ‘Kaggle Grandmaster’ to each firm on the planet. That means, an organization might deal with innovation – what they wish to obtain in house exploration, scientific discovery, genome sequencing, environmental effectivity, healthcare, and so forth. AIs would offer the required machine studying to inform us how to do that notice.”

“Every so often a expertise comes alongside that not solely solves an issue for as we speak’s challenges, however transforms your complete business and takes it into the longer term,” Saurabh concluded. “NEO will do this for ML engineering. We’re on the cusp of one thing greater than we will think about.”

About NEO

NEO is a multi-agent system able to fixing complicated machine studying issues – from knowledge engineering to ML mannequin implementation – lowering the heavy lifting of ML engineers, with human steerage. The aim is that each firm on the planet can have a best-in-class AI professional on its crew, accelerating scientific breakthroughs for all humanity. Two brothers – Saurabh and Gaurav Vij – with a mixed eight years of expertise in machine studying lead the group, now in stealth mode.

For extra info, go to: heyneo.so.

Comply with us on social media:
X: @metneo
Introductory message: NEO AI
LinkedIn: Hey-neo

Media contact:

Erica Zeidenberg
Scorching tomato advertising
[email protected]
925.518.8159

SOURCE: NEO