Recently, we announced that Softacom is embarking on a new journey in AI. At our Softacom AI Lab, we use neural networks and machine learning technologies to enhance computer vision and video processing. Previously, our Founder and CEO, Serge Pilko answered popular questions about Softacom AI Lab, how it was created, and how it operates.
In this article, Serge Pilko shares his vision for the future of Softacom AI Lab and gives recommendations to organizations that are only considering AI implementation into their products.
What are the Plans for the Near Future of Softacom AI Lab?
I won’t be modest here. Our plans are ambitious, and they often stem from the new challenges we face while implementing projects for our clients. The Softacom AI Lab is designed for visual interaction with neural networks across all stages of their lifecycle, from modeling to deployment in the final product. Many market products, solutions, and startups focus on specific areas, like data annotation tools or transformers. We aim to cover the entire process from A to Z within a single desktop solution.
Our approach combines our proprietary algorithms and solutions with open-source or commercial tools, integrating them into an environment.
Currently, we are planning to develop the possibility of both working locally (in particular, training) and transferring the model and dataset to remote computing resources based on RACK architecture – a kind of automation of DeepML Ops. Besides, important directions include the development of transfer learning path functionality that will display the full training history of models and changes in results, as well as a system for automating user assistance for annotating data utilizing other trained neural networks.
What Inspires You to Work with AI Technologies?
It’s fascinating and non-trivial. It’s what motivates our team. It’s the future.
Why Should Companies Choose Softacom to Implement AI Solutions?
I’d like to point out that companies should choose us only if we are the right fit for their specific task. Fortunately – or unfortunately – we can’t do everything for everyone. The AI development world is huge. We are working in areas such as neural network training, and computer vision, and integrating these with Generative AI into unified solutions.
As for “why choose us…”: because we quite often ask our clients “why” regarding their product, project, or business. Technology and implementation are only half the battle. Without a deep understanding of the goals and value a product must bring, from our point of view, it’s impossible to create a wow solution.
What Would you Recommend to Companies that are Just Starting to Explore AI?
Here are my thoughts, which might be useful for companies looking to implement AI solutions for their B2B or B2C products. I’ll exclude companies like ours that specialize in development. Let’s create a list of recommendations. These are the questions companies should ask themselves (without prioritizing their importance):
- Has anyone successfully implemented a similar solution? Is it technologically possible?
- Can we measure how this AI solution will impact our product’s financial or functional metrics?
- Will it be profitable? Or do we need to spend $100,000 on computational resources to test a hypothesis?
- Is it worth investing in an internal AI department, or should we outsource this subproject?
In summary, I believe tech companies should invest at least in basic AI knowledge within their organization, because it’s not just a future trend, it’s already a technological standard today.
