0:04 my teams have worked with or advised
0:06 many companies both large and small on
0:08 building a large variety of geni
0:10 applications I'd like to share with you
0:13 in this video some of the best practices
0:15 I'm seeing as well as what the typical
0:17 team to get started on such project
0:19 might look like the most common roles
0:23 for building gentv applications would be
0:25 a software engineer who would be
0:27 responsible for writing the software
0:29 application and making sure that it runs robust
0:30 robust
0:32 and I've seen that when a software
0:35 engineer puts in just a bit of effort to
0:37 learn at least the basics of large
0:39 language models in prompting then they
0:42 can be very effective in a small team
0:46 building L based applications so if
0:48 you're on a team that already has some
0:50 software Engineers it might be worth
0:52 encouraging them to consider spending
0:54 just a little bit of time to learn at
0:57 least the basics of Ls and prompting a
1:00 second row that is quite common on teams
1:02 to build applications would be the
1:05 machine learning engineer and machine
1:06 learning Engineers are typically
1:09 responsible for implementing the AI
1:11 system many machine learning Engineers
1:13 have been building AI systems even
1:16 before generative AI took off and I
1:18 found that a machine learning engineer
1:20 that spends just a bit of effort to
1:23 learn about LMS and ideally not just
1:25 prompting but some of the more Advanced
1:27 Techniques like Rag and fine-tuning such
1:30 a person can be very effective playing a
1:34 role building applications and finally
1:37 one other role that I see in some teams
1:38 but less common than the software
1:40 engineer and machine learning engineer
1:42 would be the product manager and they
1:44 would be the person with primary
1:45 responsibility for identifying and
1:48 scoping the project and making sure that
1:50 whatever is built is useful for
1:53 customers lastly how about the promt
1:55 engineer role there's been a bit of
1:57 media hype about this role but what I'm
1:59 seeing is that very few companies are
2:01 hiring this as a dedicated role what
2:02 happened was a small number of companies
2:04 advertised a small number of job
2:06 openings for very well-paid prom
2:08 engineers and this generated a lot of
2:10 media hype that someone could make a lot
2:13 of money by prompting but if you look at
2:15 the actual job description of prom
2:18 Engineers promt engineering jobs
2:20 actually require doing a lot of other
2:22 tasks than just writing prompts and they
2:23 actually look more like machine learning
2:25 Engineers that have additionally learned
2:28 to prompt so don't bind the hype of the
2:30 prompt engineer Ro
2:32 what actually happens in practice is
2:34 that most companies are counting on
2:36 machine learning Engineers that have
2:39 also learned LS or learned prompting and
2:42 is actually not that easy to get a job
2:44 and nor company is hiring that many
2:46 people whose only job is the right
2:48 prompts if you're building an L based
2:50 application it's often possible to get
2:53 started with a pretty small team so I
2:55 definitely see companies start to
2:57 experiment with even just a one person
2:59 team such as a software engineer
3:01 engineer who's learned some prompting or
3:03 a machine learning engineer who's
3:06 learned a bit about prompting in alms or
3:08 maybe you could just start by yourself
3:11 by experimenting and prototyping using
3:13 some of the web interfaces to try to get
3:15 a sense of what might be feasible I do
3:18 see a lot of TW person teams as well and
3:21 if you have two persons in a team
3:23 probably the most common configuration
3:25 is a machine learning engineer plus a
3:28 software engineer but I've seen many
3:31 other configurations also work well such
3:33 as a software engineer who's learn
3:35 prompting and a product manager or
3:38 really to generally enthusiastic people
3:40 that know a bit about how to write
3:41 software and are willing to learn how to
3:44 use these tools to build new and exciting
3:45 exciting
3:47 applications sometimes for the larger
3:49 teams you also see some other roles like
3:51 data engineer data scientist project
3:53 manager or machine learning researcher
3:54 let me quickly talk over these roles as
3:57 well in case you see them in the company
3:59 so data engineer is usually responsible
4:02 for organizing the data and ensuring
4:04 data quality and often also the security
4:07 of the data and data scientist is
4:09 usually responsible for analyzing data
4:11 to make recommendations to guide project
4:14 or business decisions project manager
4:15 can be responsible for coordinating
4:18 project execution and machine learning
4:20 researchers are usually responsible for
4:22 developing Advanced AI Technologies or
4:24 adapting Advanced AI Technologies to the
4:28 particulars of your business so G of AI
4:30 has lowered the cost lower the barrier
4:33 to entry to building AI based
4:35 applications if you or your team have an
4:37 idea I'd encourage you to try to find a
4:39 resources to prototype and try something
4:41 out and see if you can build something
4:44 for yourself or for your business before
4:47 we wrap up the section on gen of AI and
4:50 businesses I'd like to go through an
4:53 analysis of how AI is affecting
4:54 different job roles as well as different
4:57 industry sectors let's take a look at
