FedML Secures USD 11.5M to Drive Collaborative AI Model Training at the Edge

Highlights:

• The collaborative approach of FedML has the additional benefit of lowering the high resource requirements of AI training.
• FedML eliminates the requirement for purchasing pricey graphics processing units by enabling customers to train and serve their models on any hardware.

FedML Inc., a startup focused on collaborative machine learning, announced raising USD 11.5 million in the seed investment round.

Two distinct tranches made up the round. While the second brought up USD 7.2 million and closed earlier this month, the first raised USD 4.3 million and was announced in March. Road Capital, Sparkle Ventures, PrimeSet, Finality Capital Partners, LDV Partners, AimTop Ventures, Robot Ventures, Wisemont Capital, Modular Capital, and the University of Southern California were also investors in the round, which Camford Capital led.

By sharing data, models, and compute resources, FedML has developed a collaborative ML operations platform that enables businesses and developers to collaborate on machine learning projects. The company claims that its objective is to construct an ecosystem for the cooperative development of unique AI models.

The startup explains that to improve the ability to handle customer service, product design, and business automation tasks, many businesses are interested in training and fine-tuning AI models on their private datasets. However, with current cloud-based AI training methods, it is challenging for them to securely utilize this data, which is frequently controlled or compartmentalized.

A federated machine learning platform called FedML’s solution enables the cooperative edge training of AI models on private and segregated data. It implies that moving data to a different location is not necessary. A healthcare institution might develop AI models capable of detecting uncommon genetic illnesses by training it on extremely sensitive data maintained at several hospitals due to this “learning without sharing” approach.

The collaborative approach of FedML has the additional benefit of lowering the high resource requirements of AI training. It is well known that ChatGPT’s developer, OpenAI LP, invested millions of dollars in its training.

Not all businesses have the same types of financial resources. FedML eliminates the requirement for purchasing pricey graphics processing units by enabling customers to train and serve their models on any hardware.

Salman Avestimehr, Co-founder and CEO of FedML, said businesses should construct their own unique AI models rather than using generic large language models produced by companies like OpenAI and Google LLC.

“Large-scale AI is unlocking new possibilities and driving innovation across industries, from language and vision to robotics and reasoning. At the same time, businesses have serious and legitimate concerns about data privacy, intellectual property, and development costs. All of these point to the need for custom AI models as the best path forward,” he added.

Since its March 2022 launch, FedML has advanced significantly and amassed a lively open-source community of over 3,000 users. Its open-source federated machine learning library has surpassed Google’s TensorFlow Federated in popularity and has completed over 8,500 AI training jobs across 10,000 edge devices. The business claims to have landed over ten enterprise contracts in the mobility, healthcare, retail, and financial services sectors.

The FedLLM, a tailored training pipeline for creating domain-specific LLMs on confidential data, was its most recent innovation. It is rumored to work with well-known LLM libraries like HuggingFace and DeepSpeed, and programmers may get going immediately by adding just a few lines of source code to their programs. Next, FedLLM handles all problematic processes required for training, serving, and monitoring client LLM models.

Ali Farahanchi, Camford Capital Partner, said, “FedML has a compelling vision and unique technology to enable open, collaborative AI at scale. In a world where every company needs to harness AI, we believe FedML will power both company and community innovation that democratizes AI adoption.”