WEBVTT 1 00:00:00.690 --> 00:00:04.010 I'm very happy to be here in London today with you, Chris. Chris, 2 00:00:04.140 --> 00:00:07.860 you're one of BCG X managing directors and partners. 3 00:00:07.980 --> 00:00:09.150 Recently elected. 4 00:00:09.570 --> 00:00:14.070 You are leading all our AI offers into health care 5 00:00:14.520 --> 00:00:18.450 R&D, very cutting edge, very cutting edge domain for us. 6 00:00:18.780 --> 00:00:21.780 You are a scientist at heart and at core. 7 00:00:22.230 --> 00:00:26.850 You did a PhD at Oxford in the med school there. So to start with, 8 00:00:27.330 --> 00:00:31.920 how can one decide to start a PhD in the Oxford Med 9 00:00:31.920 --> 00:00:32.753 School? 10 00:00:33.390 --> 00:00:35.220 Well, I was always interested in science. 11 00:00:35.550 --> 00:00:38.340 I've always liked biology and chemistry in particular. 12 00:00:38.670 --> 00:00:41.370 And so I've always wanted to work in that field. 13 00:00:42.120 --> 00:00:45.180 I'd done a first degree in that space and then I thought, this is interesting, 14 00:00:45.180 --> 00:00:47.970 I want to do more. And so did my PhD then. 15 00:00:48.210 --> 00:00:53.130 And the interesting thing is we then at the time were studying something which 16 00:00:53.130 --> 00:00:55.800 was quite novel at the time, which is coronaviruses. Now, 17 00:00:55.800 --> 00:00:56.790 this is a long time ago. 18 00:00:56.790 --> 00:01:00.690 This is a good 10 years before the big pandemic where we did some of the early 19 00:01:00.690 --> 00:01:02.880 work there. But really exciting, really cutting edge. 20 00:01:03.120 --> 00:01:08.010 And any breakthrough discovery insight that you had during your time 21 00:01:08.010 --> 00:01:09.030 as a researcher? 22 00:01:09.540 --> 00:01:12.540 So we actually looked at how coronaviruses evolve. 23 00:01:13.410 --> 00:01:16.770 They have a tendency to jump from one species to the next. 24 00:01:16.770 --> 00:01:20.130 And so we looked at some of the genetic and molecular mechanisms for that. 25 00:01:20.250 --> 00:01:22.560 So some quite interesting insights there of how that happens. 26 00:01:22.980 --> 00:01:26.520 So you started with a few years in hardcore research and academia, 27 00:01:26.520 --> 00:01:31.230 then you decided to move to business first, UCB and then BCG X. 28 00:01:31.590 --> 00:01:32.423 Why? 29 00:01:32.820 --> 00:01:36.360 Well, I've always liked the applied aspect of the scientific work, 30 00:01:36.600 --> 00:01:39.300 finding out new things, but then making them useful, making them, 31 00:01:39.300 --> 00:01:43.080 bringing them to people, discovering new drugs and bringing them out, 32 00:01:43.290 --> 00:01:47.850 those sort of things. So that's why I moved into industry. And then moving into, 33 00:01:47.880 --> 00:01:49.710 first consulting and then data science, 34 00:01:49.950 --> 00:01:52.470 I thought that was sort of the next stage in that evolution. 35 00:01:52.920 --> 00:01:57.660 And in that evolution now are the seniors within BCG X leading that super 36 00:01:57.660 --> 00:02:01.230 advanced software we have. Where do you bring science? 37 00:02:01.950 --> 00:02:04.710 I mean, science is part of what I do almost every day, right? 38 00:02:04.710 --> 00:02:08.970 I work in the field of R&D, so science is with me every day. 39 00:02:09.120 --> 00:02:11.670 We look at how do we discover drugs, 40 00:02:11.670 --> 00:02:14.700 how do we use AI to do that in a better way, a more efficient way, 41 00:02:14.700 --> 00:02:18.210 a more effective way? How do we run clinical trials? How we can do that better. 42 00:02:18.210 --> 00:02:21.210 So, I mean, science I deal with on a day-to-day basis. 43 00:02:21.480 --> 00:02:22.320 And more broadly, 44 00:02:22.320 --> 00:02:27.000 how do you see the evolution of that intersect between business on one 45 00:02:27.000 --> 00:02:29.220 side and science on the other side? 46 00:02:29.490 --> 00:02:33.060 It's a huge topic right now if you just think about AI and how that's been 47 00:02:33.060 --> 00:02:36.420 evolving in the last five years, right? I mean, those are scientific insights, 48 00:02:36.420 --> 00:02:39.420 scientific methods and techniques that are now being applied in a lot of 49 00:02:39.420 --> 00:02:43.410 different places. So yes, science is having a real impact in business. 50 00:02:43.740 --> 00:02:48.360 And I really think that time between what's in the lab and what's in the market 51 00:02:48.360 --> 00:02:51.000 is shrinking by the day. 52 00:02:51.000 --> 00:02:55.110 So I can see that super well. And speaking of research, 53 00:02:55.110 --> 00:02:59.680 you've just published an article in the clinical trial paper 54 00:03:00.880 --> 00:03:05.440 on how we can use LLMs to accelerate and manage very 55 00:03:05.440 --> 00:03:09.370 differently the clinical trials. Can you share a bit of the topic with us? 56 00:03:09.460 --> 00:03:11.050 Yeah. So during a clinical trial, 57 00:03:11.050 --> 00:03:12.880 you have to write a lot of different documents. 58 00:03:12.880 --> 00:03:15.130 You have to document exactly what you want to do, 59 00:03:15.400 --> 00:03:18.160 the results that come out of the trial. Those are very large, 60 00:03:18.160 --> 00:03:19.330 very complex documents. 61 00:03:19.360 --> 00:03:23.890 And what we looked at is how can you use large language models to do that 62 00:03:23.950 --> 00:03:24.730 better? 63 00:03:24.730 --> 00:03:29.260 And basically what we find is that if you just take an off-the-shelf tool, 64 00:03:29.260 --> 00:03:33.700 like a charChatGPT, it works, but it's not perfect. 65 00:03:33.880 --> 00:03:37.030 And what we then show is that if you engineer those models a little bit, 66 00:03:37.030 --> 00:03:40.270 you feed a little bit of information in and you optimize a little bit, 67 00:03:40.270 --> 00:03:42.190 then the quality of the writing becomes much, much, 68 00:03:42.190 --> 00:03:45.550 much better and becomes useful for actual clinical trials. 69 00:03:45.670 --> 00:03:46.930 That was the publication. 70 00:03:47.830 --> 00:03:51.640 And Chris, I know you have a mission at heart, not just for you, 71 00:03:51.670 --> 00:03:56.170 not just for BCG X, but for the world at large. Can you share this one? 72 00:03:56.410 --> 00:03:57.243 Yeah. I mean, look, 73 00:03:57.580 --> 00:04:01.810 I want to help bring new medicines out into the world and to patients. 74 00:04:02.020 --> 00:04:05.500 I want to do that in an innovative way using some exciting AI and other 75 00:04:05.500 --> 00:04:08.500 techniques, right? That's why I'm here. And that's what we do every day. 76 00:04:08.830 --> 00:04:12.130 And that's where you start to have a massive impact with the work you do. 77 00:04:12.130 --> 00:04:14.650 So congrats for that. Thank you for sharing, Chris.