Meet Francesco Farina: Using tech to optimise portfolios, increase investor efficiency and support founders

Following his piece on designing the optimal early-stage venture portfolio and the launch of our Portfolio Simulator, today we're interviewing Francesco, one of our senior machine learning engineers and a member of the team behind the work.

Get to know him, his vision for technology and investment, and how he’s using his deep understanding of machine learning to bolster our data-driven strategy. You can read his bio here.

What brought you to venture capital?

I’ve been investing and trading since I was 18. During my years as a researcher – working on many topics, including distributed systems, optimisation, machine learning and control theory – and while working at GSK, I got more and more interested in how these technologies could be applied across multiple fields. So when I came into contact with venture capital, it looked like the perfect fit. I’d get to dig into new technologies and understand how they’ll impact our future – and I’d be able to invest in and help the companies building them. It was everything in one place.

How are you using technology at Moonfire?

Basically every task you have at a VC firm, you can enhance and scale it with data or technology in some way. We’re trying to build tools that help us find the right founders faster, and build tools for those founders when they join Moonfire.

And with our recent work on portfolio construction, that’s useful for us in thinking about how we approach our own portfolio strategy, but we also think about how these tools can create a better VC community. We can share what we find and let everyone benefit from it.

What are you and the team working on right now?

After spending the last year focused on building out our automated sourcing, screening, and evaluation engine, the team is currently focusing on two main things: investor efficiency and portfolio support. We want to use technology to make investors more efficient and accurate, so they’re spending more time doing the things that they actually like and making better decisions.

This is necessary because as I mentioned, we have spent a lot of time building technology for sourcing and evaluating deals, and we have become quite good at that. We’re now able to see two million companies a year instead of 10,000, and we can filter them in a pretty efficient way so that our investors only have to deal with the most relevant ones. I’m also working on trying to make our decision-making more accurate, more consistent, so every deal we look at is treated in the same way and we’re making decisions in a fair, non-biased way.

We’re also building tools for our founders. We’re planning to release one that will help founders in our portfolio find possible investors for their next fundraise based on what the investors have invested in in the past, their interests, their typical cheque size – these sorts of things. We want to make getting to the next part of their journey a bit easier.

There’s a lot of other ideas for applications and platforms for founders on the table, but that’s for another day!

What other problems are you keen to solve?

There are some big open questions in VC.

We’ve started to answer the one around portfolio construction – what’s the optimal way to build your portfolio? – particularly for pre-seed and seed stage. But it’s not complete. If we change some of the core assumptions of our fund strategy, the argument we have is completely different, and we’ll have to build a different model. This is a problem I’m still super interested in and want to dig deeper into.

I also think technology should help in making VCs more proactive rather than reactive. When we started as an industry, early, mid last century, it was really a proactive thing. You went to look for the new breakthrough technology and help the founders build a product from scratch. But it’s becoming more reactive, and more prone to falling into hype cycles.

We’re seeing it now with generative AI. Of course, there are some incredible products coming out, and the technology is becoming more accessible so everyone can play with it and start building stuff with it. But we’ve been discussing these topics for the last 20 years. And there’s a danger of assigning strengths to this technology that it doesn't yet have.

I think there's more to do in monitoring actual new technologies and models in this space, and this is an interesting direction we could take a sourcing and evaluation model. By monitoring the latest technologies, which is something that most people aren’t doing right now, unless you’re in an incubator from a university, you can find these opportunities earlier and jump on them if they align with your thesis.

So will we see a rise in technical VCs?

I think so. New technologies are becoming more and more complex to understand, and to understand whether it’s just hype or if it’s something new, something that will grow, you need to understand that technology very well. So with generative AI or large language models, for example, I think you need to really understand what they are and what they’re capable of in order to understand their limits and where they can go, before writing a thesis around them and trying to predict what will happen.

And it’s something I often hear from founders; they’re really surprised when they talk to us because they know that we can deeply understand what they’re building. So we’re talking with a startup right now that’s building an advanced AI-based product, and they were struggling to get the message through with some other VCs because they couldn’t understand how difficult the problem is and how potentially revolutionary the technology can be. Because we as a team have a background in machine learning, we can ask relevant questions, suggest different ideas – it’s as it happens in research. It’s about discussing and coming up with new ideas, and deeply understanding what’s being built and why, and they feel that they can get that when they talk to us. Even companies that we’ve already invested in, I think that’s still something they find very attractive.

What are you most excited about at the moment?

I think AI is the key to solve most of the challenges we have in our time.

We can leverage massive amounts of data to progress in ways we just couldn’t a few years ago. We can use them together with the laws of nature we discovered over the years to progress fundamental science at an extraordinary pace. And all of that is, and will be, massively reflected in our lives: think about the marriage between AI and biochemistry and its potential. I wouldn’t be surprised if in 50 years most of the diseases we know today will be just a distant memory. All in all, I think it’s an exciting time to live in.

What do you think VC will look like in 2030?

Technologists and scientists will have a bigger role in the industry, with the rise of bio and further advances in AI. Everyone will be using more data and technology in some form or other, and machine-aided decision making. Predicting trends will probably still be an investor’s prerogative, though. At least, I hope so. It would be sad if a machine was making all the decisions.