As we gear up for our San Jose Summit, we're touching base with yet another of our speakers.

Vahan Petrosyan, CEO and co-founder at SuperAnnotate, will be in San Jose with his session, "Image annotation is dead: Semi-supervised active one-shot is the future".

As the only company in the market entirely dedicated to solving the data problem for machine learning, SuperAnnotate democratizes access to AI innovation.

📍Headliners Stage

⏰ Day 1 [April 16] 11:15 - 11:40 | GMT -7

🗺 San Jose Marriott, 301 South Market Street, San Jose

1. Tell us a bit about SuperAnnotate and the products and services you offer.

So, I was doing my PhD at KTH - Royal Institute of Technology, focusing on Computer Vision and Image Segmentation when I created a new algorithm that could make the creation of datasets for segmentation much easier.

Together with my brother Tigran, who was doing his PhD in Switzerland, we realized that this algorithm could provide a lot of value to everyone building CV algorithms and we decided to start a company around it. 

Now years later we have expanded well beyond that initial research but our core product is still our platform, which is built to accelerate the pace of dataset creation and model development to make large-scale data annotation projects as efficient as possible, and as of late, to empower individuals to accomplish more by leveraging advanced AI tools that allow you to build new datasets quickly. 

During the last year, we have also leaned heavily into the LLM sector. LLM and GenAI require large amounts of data for fine-tuning, often created by qualified individuals, and we realized that there was no good platform for this.

We have engaged in LLM dataset creation projects with several of the leading foundational model companies and have developed our LLM platform. Now we are focused on bringing this tooling which is now what the big pros use, to the hands of all companies wanting to leverage and fine-tune LLMs to ensure that it is as efficient as possible. 

We have also helped a lot of companies with setting up and managing their annotation projects and based on those learnings, we now help enterprises with fully managed annotation projects and in-depth dataset creation consulting services. 

2. What sets SuperAnnotate apart from the competition?

We are the only company in the market that is entirely solving the data problem for machine learning.

The approach we took is very unique for the industry. Instead of giving some tool for annotation, data curation, or model evaluation, we enable our customers to quickly build the interfaces they like, compare and visualize any metric that they are interested in, directly connect to their data sources, and fully automate their data pipelines.

3. How do your products and services benefit the AI community?

Our products and services benefit the AI community by democratizing access to AI innovation. We achieve this by simplifying and accelerating the dataset creation process, often the most challenging and labor-intensive part of developing machine learning models.

Creating high-quality datasets is crucial for training models, whether it is within computer vision, natural language, or Generative AI. However, the process of creating datasets is incredibly labor intensive, making it a costly and time-consuming endeavor.

By eliminating this bottleneck, we open the door for a wider range of creators to participate in AI innovation. No longer is the field restricted to well-funded companies or startups with significant resources. Our approach levels the playing field, allowing anyone with a transformative idea the opportunity to realize it, potentially changing the world. 

Our commitment to making dataset creation more accessible and efficient fosters innovation within the AI community and encourages a more inclusive and diverse exploration of potential AI applications. This, in turn, can lead to the development of novel solutions that address a wide range of global challenges, propelling the entire field of artificial intelligence.

4. What is the biggest challenge your company is facing?

Scaling to meet demand and bringing the pro infrastructure for LLM data to the masses. Over the last 6 months, we have seen more than 10x the appetite for using our platform to create datasets for LLMs. As a B2B infrastructure company, scaling is an interesting challenge, especially for us, where we never want to compromise the quality during that process.   

5. What can the audience expect from your session at the event?

A deep dive into the next big thing in dataset creation for computer vision. There have been a lot of really exciting advances in foundational models and generative AI that we are leveraging to build a system that allows one person or a really small team to do the work that previously required large teams of annotators both to create datasets and even to build state of the art ML models using an interactive learning environment. 

Interested in Vahan's talk? Get your tickets below:

AIAI San Jose 2024 | Deploy & scale the state-of-the-art
On April 16 & 17, join 500+ leading builders & execs bridging the gap between cutting edge research & value-driving application at San Jose Marriott.