YT YouTube2Text

YouTube2Text → Transcripts → 日本のFDEがぶつかる壁とは？デプロイメントストラテジスト経験者と語るFDEモデル浸透の鍵／LayerX／DS／コンサルタント／中村龍矢

日本のFDEがぶつかる壁とは？デプロイメントストラテジスト経験者と語るFDEモデル浸透の鍵／LayerX／DS／コンサルタント／中村龍矢 — Transcript

by LayerX 公式 · 2,742 words · 413 segments · language ja · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:06Like a lot of FDE is actually
0:08discovering the true problem, suggesting
0:11a solution that was not the most obvious
0:14one.
0:18>> I think Japan is the second largest
0:20enterprise software market in the world
0:22after the US. There's a lot of good
0:23talent.
0:28I think there are some key differences
0:29in the method of approach uh which I've
0:32seen from SI companies I've worked with
0:35here to there's a risk that some of that
0:37desire to please the customer and do
0:39everything that they say like to to the
0:42letter
0:43>> actually is not perfect FD practice but
0:46that could actually solve the problem
0:48not just what you asked to build or
0:51change
1:11deployment strategy.
1:48There are personal reasons. I really
1:49enjoy the lifestyle here. M I think that
1:52the quality of life is much higher for
1:54the same sort of cost and you have so
1:57many opportunities both to get to other
1:59parts of of Asia and within Japan. Um
2:03and that has benefits when starting a
2:04startup. Uh you can hire very high
2:08quality people uh quite inexpensively uh
2:11compared to other places. Uh it attracts
2:14global talent actually because a lot of
2:16people from Europe and America and
2:18everywhere as well as Japanese engineers
2:20choose to live in in Tokyo. Starting a
2:23business here in the AI sector, we've
2:25actually found there's a lot of good
2:27talent.
2:28>> It's actually a great market. Uh I think
2:30Japan is the second largest enterprise
2:32software market in the world after the
2:34US. So there's a lot of customers, big
2:36customers,
2:37>> uh who want, you know, uh digital
2:40transformation, who want to use
2:41software. There's a big emphasis on
2:43increasing efficiency and then but also
2:45the challenge you know people are still
2:46using a lot of older technology there's
2:49a gap in adopting a lot of u newer
2:54technology and so trying to solve that
2:55is both a great opportunity and an
2:57interesting challenge I More.
3:19Yeah, it's interesting. I think suddenly
3:21uh the word became very popular try and
3:23introduce this role and use it in their
3:25companies. I think it's it's like many
3:27things there is a risk that it's just
3:30you becomes another buzz word and so
3:33people can say I'm doing FD but they
3:36just mean your engineer goes to the
3:39client's office whereas really it's a
3:41whole mindset and way of uh and way of
3:45working. M what your thoughts on other
3:48company trying to um learn from parent
3:51and trying to reproduce Palanteer's
3:53successful success.
3:55>> It's great that people are picking up on
3:57this finally and um I can see why that's
4:00happening now. Uh I think that as the
4:05speed of building software or at least
4:07the perception of the speed of building
4:09software has become so fast,
4:13>> customers are more and more sensitive to
4:16wanting not one sizefits-all
4:19solutions uh but something that really
4:21solves their uh specific problem.
4:24Deployment strategist
4:28>> deployment strategist is the
4:30complimentary role to uh FD for deployed
4:34engineer.
4:34>> Uh they were often referred to as echo
4:36and delta uh where echko was the
4:38deployment strategist. It's not exactly
4:41like
4:42any other role.
4:44>> Um and honestly it would depend on the
4:46person. people were hired for deployment
4:48strategist role normally a little bit
4:50later in life like many FTEEs might be
4:53fresh out of uh college 22 23 years old
4:57whereas deployment strategist could be
4:59often more than 25 up to 30s kind of
5:02range even when starting and it it would
5:05span sort of everything from solution
5:07architect through to consulting um
5:11through sometimes to elements of data
5:14science and all the way to sales But um
5:18I think it's maybe it's better to give
5:19an example of you know how a deployment
5:21strategist works and how they work with
5:23um for deployed engineers. Your leverage
5:26person saleserson would introduce you to
5:28a client and they say you know we've
5:30picked Declan to be the the leader on
5:33this project here and give a rough idea
5:36of the problem and introduce you to
5:38whoever the the point of contact the
5:40client side was. Then it's kind of on
5:42the deployment strategist at that stage
5:45to arrange okay who do I need to meet
5:48with next? What are we going to discuss?
5:50How do we do the scoping to discover the
5:52problem and work out how our technology
5:56awesome new technology we could build
5:58will solve this. So really to get to the
6:00core of what is the problem, what are
6:02the processes, do that analysis, then
6:05translate that to something that we
6:07could build.
6:14a very big variety of the deployment
6:16strategists. I think like I said it's
6:18people tend to start later. The reason
6:19being that
6:22you can tell someone's you know great
6:24intelligence or ability to write good
6:26code or pass tests
6:29>> when they're much younger. But generally
6:32for deployment strategists to be good at
6:35it you have to have some experience in
6:36the world. You have to have shown that
6:38you created something or led something
6:42important and can just sort of survive
6:45[laughter] a few hard knocks in business
6:47maybe. So we had a lot of variety of
6:48people. Some of my favorite deployment
6:51strategists or people I thought were
6:52great actually came from an electrical
6:54engineering background which I think is
6:56they have a different way of thinking
6:58than a software engineer but still very
7:00logical and problem solving oriented uh
7:04but about something physical. However,
7:07you know, we had people who were
7:08ex-founders who'd founded a company and
7:11left or sold it or it had failed but
7:14wanted to get up and try again. We had
7:16people who came from um an academic
7:19background
7:20>> uh but normally someone who's like tried
7:22to do something entrepreneurial within
7:24that. So there were some people from
7:26consultancies
7:27>> like uh you know Accenture, Deote, um
7:30PWC etc. And I think many of those
7:33people struggled actually those
7:35companies do a lot of impressive stuff
7:37but they are not as focused on creating
7:40something that does action that changes
7:41the world that makes important decisions
7:43for people or helps them to make
7:44important decisions and the idea of
7:46building a product I think is a
7:48different mindset than most of those uh
7:50consulties. So some of those people
7:52really struggled
8:05So, Palantia Japan had actually launched
8:08previously and had uh an earlier
8:11>> version. They had launched with So done
8:14a number of projects. Um and they
8:17continued but during the coronairus
8:19crisis uh they slowed down a lot and
8:22reduced size uh and there was
8:24essentially like a relaunch um after the
8:29borders reopened and as part of that uh
8:31I was brought in. I had been building
8:34incubators we called them in the UK uh
8:37for healthcare which were trying to
8:40rapidly discover small scale new
8:44projects, new uh I guess products we
8:46could build on top of the Palanteer
8:48software to help in healthcare uh
8:51settings and to grow those into their
8:54own small [music] um enterprises. So for
8:57Japan, they wanted to take a lot of
9:00small uh contracts and small
9:03relationships and find a way to grow
9:06those into like a large uh ecosystem, a
9:09large number of those. From my
9:10experience, I was asked to help out with
9:12that and took on this uh slightly weird
9:14role of head of new ventures
9:16>> um which was about that like 0 to1
9:19growth with a larger variety. It ended
9:22up changing a bit because we then
9:23launched AIP. You know, the the release
9:26of um I think at that time chat GBT4
9:30revolutionized what was possible to do
9:33especially with like smaller scale data.
9:34You know, we'd always focused on how do
9:37you manage very very large data sets
9:40quickly and uh combine them together.
9:43But we're missing
9:45how to automate a lot of processes that
9:49uh require, you know, handling
9:51documents. There's only a few hundred.
9:52There's only a couple dozen, but they've
9:54got a lot of information in them.
9:55>> Typically, this is a thing humans still
9:57had to do. And so, the introduction of
9:59these technologies kind of
10:00revolutionized what you could do. We
10:01pivoted around it. We started making a
10:03new product about it. and I started
10:04helping to introduce that product here
10:07and particularly working with some of
10:08our partners like Fujitsu to uh enable
10:11them to like become a a delivery partner
10:13at like at scale uh and resell globally.
10:20>> I think there are some key differences
10:22in the method of approach and in how
10:26often and um how much you push for
10:28things. So I mentioned sometimes there's
10:31got to be a little tension. You should
10:32be looking for things that are maybe not
10:35exactly what you were asked, but that
10:36could actually solve the problem, not
10:38just what you were asked to build or
10:41change. Because of the way Japanese
10:44companies tend to work with system
10:46integrators and software partners in the
10:48past,
10:49>> they're not used to that kind of push
10:50back. It has to be done sensitively.
10:52Maybe you have to pick the right person
10:55whereas an FTE would directly say this.
10:57Uh in other places the account owner
11:00like presents a whole uh more formalized
11:04set of suggestions. Also we would always
11:06empi emphasize you know going on site
11:08just just go there just be in the office
11:10and like talk to anyone who who you can
11:13and find stuff out. I think the office
11:14culture here is not like that and often
11:17there are even laws um about
11:21how they can work with contractors and
11:23so on. So, [music]
11:25you can't just send somebody to go and
11:26sit on the doorstep until they're
11:28allowed in.
11:29>> We have to do more remote things. We
11:30have to create the like the more formal
11:35set of steer codes uh and weekly
11:39check-ins and other types of of like
11:41ritual with the customer at the start of
11:43the the deployment.
11:48for layer X. Um I've been advising on uh
11:52how to train FTEEs, uh how to develop as
11:55an FDE, the way of working between uh DS
11:59and uh FTE in particular. Uh so yes
12:02we've done a number of workshops um
12:05[snorts] focusing on I think certain
12:08techniques mostly so specific examples
12:12and specific techniques
12:14>> uh both for individual um work as a DS
12:18or FTE and as teams. So for instance uh
12:21we did a whole workshop on you know what
12:23a murder board is which is this
12:24technique for preparing under pressure
12:27for very important um client
12:30interactions. We spent time on how to
12:34find KPIs that are really impactful and
12:38how to present those to customers to uh
12:41grow projects or find new areas of work.
12:43And we talked about a lot of the uh the
12:47way that uh DS and FD can work together
12:50to maximize the the impact of
12:54discovering client problems and rapidly
12:56iterating on solving them.
13:02I was uh really impressed with um the
13:05kind of questions that people asked me
13:07clearly like thinking a lot about how
13:09can they go beyond uh just doing exactly
13:12what the customer asks to
13:15>> uh actually grow the size of their
13:17impact. So like a lot of the right kind
13:19of thinking like that. I also uh found
13:22like when I was asking people to reflect
13:24on [snorts] what's an example from your
13:26day-to-day work of um how do you for
13:31instance deal with customers question
13:33you about why don't I just build this uh
13:35myself in claude code or co-pilot like
13:39show a lot of uh introspection um
13:41already about that sort of thing. So I
13:44think the like the raw talent of the
13:46team is is good. Um I'm hoping that some
13:50of the techniques that I provide uh or
13:52experience that I share can help them
13:54think about mostly think about how to
13:56just share that with each other.
13:58>> I think sometimes people had not thought
14:00about just sharing everything that they
14:03learn so that everyone else can benefit
14:05from it. So individuals are getting
14:06great learning and experience. Um, and
14:10if your whole team can actually share
14:11all of that stuff together, then they're
14:13just going to continue to uh be like one
14:16great unit rather than a bunch of very
14:19strong individuals.
14:26>> I think it's those two similar to what I
14:27just said. um like building the habit
14:30and the muscle of sharing everything
14:32whether that's just through
14:35shared documents and and emails where
14:37everyone says what they've done and what
14:39they've learned and particularly if
14:40something went bad you know not to to
14:42hide it but to say hey you know
14:44>> we had this interaction with the
14:45customer and it went badly we we
14:47overpromised and we couldn't deliver or
14:50we thought that we solved the problem
14:52doing it this way but we missed an
14:54important
14:55>> component you know those are things
14:57everyone learns from um and I think the
14:59openness of conversation there will help
15:01you a lot. I think there's also like
15:03this formalizing some of that. So just
15:06remembering to share or remembering to
15:07talk or having casual conversations is
15:09not that effective if it's not in a
15:11framework of
15:12>> these things like the murder boards like
15:14the uh things like five wise a technique
15:16where if things go wrong how you examine
15:19that and make sure it doesn't happen
15:20again
15:21>> other types of like design review and so
15:23on things that are slightly formalized
15:24that you have to do at certain periods
15:26in the project it just means you never
15:28forget
15:29>> to do that and it's all recorded so new
15:32people can pick it up as well I think
15:33those some of the key
15:34Maybe one more thing is having you know
15:38shared everyone's going to have
15:39different opinions actually of course
15:41but to come back to the point about for
15:44instance customers asking why don't I
15:46just build this in clawude or with some
15:48vibe coding and you know you guys have
15:51software that you spend a long time
15:52building and you have practices that uh
15:55make your ability to uh execute it and
15:57link everything up end to end so much
15:59more powerful and less fragile than what
16:01someone can do like that but having the
16:04whole team know what is the product
16:05vision, you know, what is the business
16:07vision and being feeling like they're
16:09part of that same mission that is so
16:12powerful and just means that everyone
16:14always knows this is why what I'm doing
16:16is the right thing and I believe in it
16:18so the customer can believe in it as
16:19well.
16:28Sometimes it sounds a little too uh
16:30grandiose, but like I really want to
16:32help grow the AI scene, I suppose, the
16:35digital enhancement scene, like the the
16:37growth of companies in Japan. I want to
16:40help to make that incredibly successful.
16:42I want to uh help make Japan a place
16:45that I can stay in and want to live in
16:47uh for a long time. and to see like a
16:50healthy culture of startups where people
16:54can feel empowered to start their own uh
16:57enterprises and empower the existing
17:00like large traditional Japanese
17:02companies
17:03>> with uh the kind of technology and
17:06talent that it might be harder for them
17:08to to get otherwise.
17:09>> So like I think that scene is starting
17:10to grow now. Um, and I love to see a lot
17:13of international talent being attracted
17:15to Japan, but also Japan's own engineers
17:18getting stronger and stronger.
17:20>> Um, you know, I really think that
17:22there's such an opportunity to be like
17:26less of just importing things from the
17:28US and being able to build uh more
17:31solutions
17:32>> here that can help Japan then sell to
17:36the rest of the world and uh grow that
17:38way.
17:41So I guess h how do you how do you
17:45envisionage what uh the expansion of AI
17:47your version of AI workforce looks like
17:49and is it different from that kind of uh
17:53very American view of completely
17:55replacing
18:01Quality of life.
18:30You've been sending FDES to work with
18:32your clients uh your customers. How do
18:36you feel that they are responding to the
18:38idea of FTE?
18:42Foreignech.
19:41You
19:54Bye-bye.

About this transcript

This page contains the full transcript of 日本のFDEがぶつかる壁とは？デプロイメントストラテジスト経験者と語るFDEモデル浸透の鍵／LayerX／DS／コンサルタント／中村龍矢 by LayerX 公式, generated from the public captions YouTube serves with the video. The transcript has 2,742 words across 413 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.