AI Technology Projects: The Case for New Skills

There are a new crop of start ups looking at disrupting knowledge-intensive areas with a mix of old and new. One that caught my eye recently is Cookr. They are building an "e-commerce marketplace that offers a wide range of healthy alternatives to restaurant-delivered food." In addition to using technology to create an alternative to Uber Eats and DoorDash etc, the article talks about what I am presuming is a machine learning model exposed through APIs (which can be self managed like commercial or non-commercially licensed LLMs like Llama or 3rd party managed like OpenAI or Gemini) to generate customized meal plans tailed to an individual's needs. Providing at scale what would be the equivalent of a nutritionist's services with home-cooked healthy meals is a perfect example of AI in action.

We are now seeing the next generation of technology investments also take these types of business 'projects' into account in addition to business and engineering applications like ERP, CRM, CAD/CAM, SCADA, etc. While traditional project management processes still apply to bring this projects to market, the technical skills needed are new (read different) from what we've traditionally seen in IT and Engineering. This includes the following:

1. ML Engineering -- understanding, leveraging and leveraging machine learning models like Large Language Model. A machine learning model serve various purposes like what you can see in the tasks list on the Hugging Face model repo. Models can be trained, tuned or enhanced with specific data using techniques like RAG working with Data Science & Engineering.
2. Data science & engineering -- clean-up/process, analyze and extract relevant data. The data in this case is what is used to fine tune the model or be used in tasks like RAG. Data engineering includes transforming data into format usable by ML models (ex: vector databases and more).
3. SW engineering -- complement the work done by ML engineering by integrating it into the broader technology ecosystem with functions ranging from crafting inputs, leveraging the out in other ways, and more.
4. Infrastructure / Platform engineering -- provide the foundation similar to what is being done for cloud native applications. I've talked about this here and here.

I am sure that start-ups like Cookr have recognized the differentiation of having AI augment their strategy and invested in these role. The hyperscalers and managed offerings from OpenAI, Azure, Google Cloud, DataBricks, AWS, etc try to simplify the amount of work in getting models up and running coupled with access to large infrastructure farms for areas like training. The article for example talks about Cookr and Gemini AI. The question everyone should ask is what is the advantage of choosing a managed model versus the trade offs on cost, lock-in, innovation and more. I always like pointing out the article by Niel Nickolaisen from issue 1.0 of the NEXT magazine, titled 'Decision Framework for Purpose Alignment'. The article discusses decision-making for the optimal use of IT resources, encouraging outsourcing services that are low on the criticality scale, but investing in those high on the criticality and differentiation scales. Embracing an open model that leveraging public or shared cloud resources and hybrid multicloud platforms (like Nutanix Cloud Platform) can help organizations start up quickly on a service provider or hyperscaler and if needed move to a hosted service provider service managed by themselves or by a system integrator partner.

System Integrators looking to expand their offerings need to invest in building out their experience in these areas if they haven't already to take advantage of the Generative AI and Machine Learning boom.

I'd recommend also reaching out to Nutanix where experts like Debojyoti (Debo) Dutta , Jason Langone , James Brown , Luke Congdon , Ron Barrett, Amit Sharma (???? ?????) , Ronnie Oomen , Gowtham Vudath , and Amalanand DSilva can help them build out offerings powered by Nutanix Enterprise AI. It can help make technology projects like an AI powered virtual Nutritionist generating customized healthy food plans a reality!