<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Andy Scheff on Medium]]></title>
        <description><![CDATA[Stories by Andy Scheff on Medium]]></description>
        <link>https://medium.com/@andyscheff?source=rss-660b8f98a41d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*7Mf5yMoKu6kX6ENii7YBEg.jpeg</url>
            <title>Stories by Andy Scheff on Medium</title>
            <link>https://medium.com/@andyscheff?source=rss-660b8f98a41d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 02:58:17 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@andyscheff/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[On Building Products with LLMs]]></title>
            <link>https://medium.com/practica/on-building-products-with-llms-b7a3204d9776?source=rss-660b8f98a41d------2</link>
            <guid isPermaLink="false">https://medium.com/p/b7a3204d9776</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[product-development]]></category>
            <dc:creator><![CDATA[Andy Scheff]]></dc:creator>
            <pubDate>Thu, 10 Aug 2023 14:42:25 GMT</pubDate>
            <atom:updated>2023-08-10T14:42:25.288Z</atom:updated>
            <content:encoded><![CDATA[<p>You have a business problem that you want to try to solve with an LLM. You try several prompting approaches but each one falls short. Infuriatingly, peppering your prompt with “please” and “thank you” does not help. Unlike broken code, there’s no clear error and there’s no clear fix. This is why you see so many listicles promising “10 ChatGPT prompts that will change your life!” Tempting as they are, these get you nowhere if what you want is to solve a complex problem specific to your product or business use case.</p><p>This was my experience. I felt despair and frustration. Building with LLMs is incredibly opaque, tedious, and uncomfortable. I resisted building LLM-powered features.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ghwLAkf3Rug9DERZSxHI3g.gif" /><figcaption>The Practica AI Challenge Coach in action</figcaption></figure><p>Now, several months into building our <a href="https://practicahq.com/learn">AI-Powered Challenge Coach</a>, I’ve developed reliable strategies to work with LLMs. I still feel the despair and frustration often, but I have tools to help me move forward. This is what I do:</p><blockquote>I still feel the despair and frustration often, but I have tools to help me move forward.</blockquote><h3>Understand the classes of problems that LLMs are great at solving</h3><p>Know how to recognize an inference problem, a classification problem, a transformation problem, or a summarization problem. Learn how to prompt to solve these classes of problems. When you have the ability to recognize these problem patterns, you’ll no longer feel like you are swinging at a piñata with a blindfold. You’ll be able to see the promising avenues to approach a complicated problem. I got my foundation from this course from <a href="http://deeplearning.ai/">DeepLearning.ai</a>: <a href="https://learn.deeplearning.ai/chatgpt-prompt-eng/lesson/1/introduction">ChatGPT Prompt Engineering for Developers</a>.</p><h3>Break down big problems</h3><p>In my experience, LLMs do not reliably solve more than one problem at a time. Most business problems worth solving are more complicated than a single inference or transformation. When you’re comfortable with the problem classes and how to solve them, use this knowledge to break down your business or product problem into smaller problems that are easier for the LLM. Solve those sub-problems independently. Your product development challenge is now to weave these multiple solutions elegantly into a single user journey or feature.</p><blockquote>When you have the ability to recognize these problem patterns, you’ll no longer feel like you are swinging at a piñata with a blindfold.</blockquote><h3>Get your whole team involved in engineering</h3><p>Finally! The interface to our engineering work is natural language instead of code. This means that non-coders can get in on the engineering action. When approaching an LLM-powered feature at Practica, we separate out the prompt engineering pieces as separate tasks, discuss what the shape of the input and output should be, and then have our non-coders work on the prompt while the engineering team builds product and infra around it. Features come together really fast this way.</p><p>All of this is high-level and hard to fully understand and implement without details and examples. I would love to expand on each point and provide case-studies from the Practica team’s experience building our AI Challenge Coach. If that’s something you want to see, please engage with this post so I know you’re out there.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b7a3204d9776" width="1" height="1" alt=""><hr><p><a href="https://medium.com/practica/on-building-products-with-llms-b7a3204d9776">On Building Products with LLMs</a> was originally published in <a href="https://medium.com/practica">Practica</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[7 Surprising Insights From the Top Engineering Career Frameworks]]></title>
            <link>https://medium.com/practica/7-surprising-insights-from-the-top-engineering-career-frameworks-66305e44e11f?source=rss-660b8f98a41d------2</link>
            <guid isPermaLink="false">https://medium.com/p/66305e44e11f</guid>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[career-development]]></category>
            <category><![CDATA[engineering-mangement]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Andy Scheff]]></dc:creator>
            <pubDate>Wed, 29 Mar 2023 15:47:36 GMT</pubDate>
            <atom:updated>2023-03-29T15:47:36.721Z</atom:updated>
            <content:encoded><![CDATA[<p>We set out on a quest to create the most comprehensive engineering career framework in the industry. Our first step was to find and analyze the top frameworks that engineering leaders spent months building. We surfaced 17 engineering career frameworks from companies like Dropbox, Etsy, and Square, and synthesized our findings into a unified framework. Then, we sent this framework to dozens of experts to get their opinions and refine <a href="https://practicahq.com/leveling-frameworks/engineering">the final product</a>. Here’s what we learned along the way.</p><h3>Most frameworks have the same common themes</h3><p>One of the core jobs of a career framework is to break down all the skills that are required in a software engineering career, across levels. Over 70 skills are used across companies (more on this below), so multiple levels of grouping are used to help navigate them. Most companies have a very similar approach that involves grouping skills into 4 high-level themes:</p><ul><li>Technical (aka technical mastery, domain expertise, craftsmanship, knowledge)</li><li>Impact (aka ownership, delivery, execution, or ‘get s*** done’)</li><li>Collaboration (aka communication, feedback, teamwork)</li><li>Leadership (aka influence, management, organizational)</li></ul><p>Some companies choose to break down their skills differently or combine two of these themes together, but the majority of teams use these themes.</p><h3>Leadership and collaboration skills dominate the frameworks</h3><p>This may not be a surprise to anyone who has tried to progress in the software engineering field: most frameworks put a far greater emphasis on collaboration and leadership skills than technical skills. This is usually clear when you look at a framework at a glance, and our in-depth research confirms it. In our comprehensive framework, only 1/4 to 1/3 of the skills are considered technical.</p><h3>The average framework has 50–75 skills, not 8–10</h3><p>Blog posts about engineering skills often say there are 10 or 12 essential engineering skills. Many job descriptions usually list out 8–10 skills. But based on our research, the average engineering framework has many more skills than that: the true number is at least 50 per company. In order to be comprehensive, we aggregated the list of skills across companies; we ultimately identified 75 individual skills that are required of most software engineers.</p><p>Some frameworks do try to shorten the list, relying on more generic terms and skill categories like ‘teamwork’ or ‘reliability’ to stand in for a long list of component skills. They do this to make the framework more understandable and less overwhelming, but it’s not optimal for learning. Identifying granular skills is the most effective way to learn and to make progress in your career.</p><blockquote>A framework should evolve over time, and it’s usually easier to add something than it is to take something away.</blockquote><h3>But, shorter and smaller frameworks can still be valuable</h3><p>Many companies publish highly detailed frameworks complete with verbose value descriptions and long lists of behavioral examples and expectations. This is not the only way to make a framework. We encountered numerous teams and companies that use concise frameworks and see good results with them. Finding the right level of detail for your team’s framework requires examining your challenges, goals, and priorities. Smaller companies, with fewer specialized teams, usually have simpler frameworks. A framework should evolve over time, and it’s usually easier to add something than it is to take something away.</p><h3>If you introduce a framework in response to a problem, you’ve waited too long</h3><p>Because frameworks are often thought of as highly detailed and precise documents, companies tend to invest a lot of time and energy in creating them. In turn, many companies put off the process of creating one for a long time. A very typical situation is that a company will create a framework as a reaction to an engineer asking a difficult question about their title or their role expectations. In this case, the company has waited too long to clearly communicate with their engineering team about role expectations and they’re likely to lose some talented engineers who see more career progression opportunities elsewhere.</p><h3>Build on top of existing company values</h3><p>Frameworks that suffer from complexity and obscurity don’t get used. In pursuit of perfection and novelty, engineering frameworks tend to introduce lots of new concepts and terminology with their framework. One easy way to boost the memorability, understandability, and thus the likelihood of success of a departmental framework is to make an effort to re-use existing company terminology. If your organization has already articulated company values or if the HR team has already defined a leveling system or compensation structure, those existing systems should be embraced and referenced throughout a framework, with department-specific skills, expectations, and levels layered on.</p><blockquote>existing systems should be embraced and referenced throughout a framework, with department-specific skills, expectations, and levels layered on.</blockquote><h3>Participation is key to buy-in and adoption</h3><p>A framework is only valuable if it’s used by individuals on the team to think about career progression and to drive productive growth discussions. Getting team members to actually adopt a framework for regular use is a challenge. One common reason why frameworks aren’t adopted is that individual engineers feel that the framework doesn’t actually map to the things they are asked to do in their day-to-day jobs. Companies that make an effort to include a broad group of engineers in the creation process, and investigate how they work and what examples of excellence look like, see stronger adoption. Furthermore, engineering organizations that have a transparent and simple process for updating a framework, or at least giving feedback on a framework, also see stronger adoption.</p><p>Thanks to the dozens of engineers, leaders, and career experts that we talked to in the course of this effort. You can view (and fork) our comprehensive framework here: <a href="https://practicahq.com/leveling-frameworks/engineering">https://practicahq.com/leveling-frameworks/engineering</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UXeXLQ-moHublTKtlOkxnQ.jpeg" /><figcaption>The Practica Engineering Skill Framework</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=66305e44e11f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/practica/7-surprising-insights-from-the-top-engineering-career-frameworks-66305e44e11f">7 Surprising Insights From the Top Engineering Career Frameworks</a> was originally published in <a href="https://medium.com/practica">Practica</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Apprenticeship as an Engineering Hiring Strategy]]></title>
            <link>https://medium.com/@andyscheff/apprenticeship-as-an-engineering-hiring-strategy-62bec84124bc?source=rss-660b8f98a41d------2</link>
            <guid isPermaLink="false">https://medium.com/p/62bec84124bc</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[hiring]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[apprenticeship]]></category>
            <category><![CDATA[engineering]]></category>
            <dc:creator><![CDATA[Andy Scheff]]></dc:creator>
            <pubDate>Mon, 06 Jan 2020 17:33:31 GMT</pubDate>
            <atom:updated>2020-01-06T20:15:05.846Z</atom:updated>
            <content:encoded><![CDATA[<p>Teaching on the job is a very simple and powerful approach to building skilled teams. Humans have used this strategy since the very beginning and continue to use it today with great results in many industries. Hiring someone intelligent and hard-working with little to no hands-on experience or special skills, then diligently and thoughtfully teaching them through applicable hands-on work is a very reliable way to produce an expert with the exact set of necessary skills. This concept can broadly be thought of as <em>apprenticeship</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*apx2jjoBIVVDDa02.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@quinoal?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Quino Al</a> on <a href="https://unsplash.com/s/photos/pottery-wheel?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>At high-growth startups, where rapid iteration is king, we don’t hire software engineers this way. We think it’s too slow, too costly, and we are skeptical of the results, so we take a different approach. We search far and wide for people who are already experts in something similar to what we need, then do vicious battle with peer companies to hire and retain them. This approach is painful and expensive, but we can’t resist the allure of a new hire that will be able to hit the ground running and immediately elevate the overall level of expertise on the team. Startup leaders soldier on searching for top talent, but at the same time, we all agree that our current hiring strategies are insufficient. We’ve lost sight of the apprenticeship approach, but it demands another look.</p><blockquote><em>We search far and wide for people who are already experts in something similar to what we need, then do vicious battle with peer companies to hire and retain them.</em></blockquote><p>Several aspects of today’s software engineering hiring market make apprenticeship a compelling option. The first, and most obvious, is the supply shortage of experienced candidates, especially non-white, non-male experienced candidates. By relaxing requirements for years of experience and/or a prestigious computer science degree and replacing those requirements with on-the-job training, we can vastly increase the available talent pool in both size and diversity. The supply shortage is a deeply-rooted, complex problem that the industry will not fully solve for a long time. Individual organizations, however, can take steps to solve the problem for themselves today by leveraging apprenticeship to draw from a larger talent pool.</p><p>In conversations with startup leaders, it’s often the second-order effects of the supply shortage that come up as most painful. Experienced engineers have tremendous leverage in the job market which can lead to selfish work habits. I often hear stories of experienced engineers prioritizing technically challenging or publicly visible work over tasks that are most impactful to the business. That same job market leverage also leads to rapid turnover, with engineers frequently hopping between companies looking for a higher salary, better equity, or more exciting work. Hiring with an expectation for growth then supplying necessary support is an antidote to these problems. It is extremely motivating to employees and it fosters deep emotional bonds and a high degree of trust within the organization. Hiring this way leads to longer, more selfless, and more sustainable employment relationships.</p><blockquote><em>Hiring with an expectation for growth then supplying necessary support is an antidote to these problems.</em></blockquote><p>Current market forces aside, software engineering is simply an ideal job for learning through apprenticeship. The ability to produce great software with modern tools is not a single skill but a collection of hundreds of skills with widely varying ramp-up time. A few of these skills take quite a while to learn or are risky without experience, but many of the most frequently needed skills are both easy to learn and safe to perform. Writing tests, integrating 3rd party services, and building clear user interface components are great examples. The prevalence of tasks that require skills like this is what makes software engineering work so well with apprenticeship. There are many ways for apprentices to deliver value to companies while they learn to take on larger, more complex, and riskier projects.</p><p>I previously led engineering hiring for an early-stage startup in SF. I experienced firsthand how startups are losing the hiring battle. Big companies maintain a strangle-hold on experienced engineers by locking them in with equity packages, bonuses, and unbeatable perks. They have greater access to new grads from top computer science programs because they can send a recruiting delegation to every top CS department and career fair. But, there are still sources of engineering talent that are overlooked by established companies (e.g. <a href="https://medium.com/@andyscheff/the-case-for-career-transitioners-b1b02dc7a51b"><em>career transitioners</em>, which I wrote about in a previous post.</a>) By embracing apprenticeship and adjusting our focus from poaching top talent to developing new talent, we can flip the script and finally build engineering teams with the market forces in our favor.</p><p><em>Thanks </em><a href="https://medium.com/u/e69c938ccf96?source=post_page-----65e9d828e56b----------------------"><em>David Whittemore</em></a><em> for helping me articulate some of the ideas in this piece.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=62bec84124bc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Case for Career Transitioners]]></title>
            <link>https://medium.com/@andyscheff/the-case-for-career-transitioners-b1b02dc7a51b?source=rss-660b8f98a41d------2</link>
            <guid isPermaLink="false">https://medium.com/p/b1b02dc7a51b</guid>
            <category><![CDATA[bootcamp]]></category>
            <category><![CDATA[apprenticeship]]></category>
            <category><![CDATA[hiring]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Andy Scheff]]></dc:creator>
            <pubDate>Mon, 02 Dec 2019 21:38:11 GMT</pubDate>
            <atom:updated>2019-12-02T21:38:11.126Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Two people writing code together" src="https://cdn-images-1.medium.com/max/1024/1*oF8SeZQINEZeb5MU4F9oNQ.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@nesabymakers?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">NESA by Makers</a> on <a href="https://unsplash.com/@nesabymakers?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>I recently drove software engineering hiring for a small startup in SF. Attracting great candidates was a tremendous challenge. A handful of behemoth companies soak up and retain a large percentage of capable engineers while startups, even well-funded ones with good traction, are left to fight over a small slice of the talent pool.</p><p>The engineering talent shortage is not new, and the industry is starting to see new sources of talent. The most prominent is a group of intensive software engineering programs (commonly known as coding boot camps). These programs churn out potential new software engineers with basic training at a rapid clip. Their graduates are, by and large, individuals who have succeeded in their prior field and decided to take a leap of faith into a completely new field with more promise for growth. These <em>career transitioners</em> are taking a calculated risk to pursue more creativity, more challenge, and more impact.</p><p>After going through a rigorous application process and grueling boot camp curriculum, they’re met by a largely unwelcoming hiring market. Most software engineering job openings explicitly require a computer science degree or state that it is strongly preferred. When I talk to engineering leaders, most say they aren’t opposed to hiring boot camp graduates but are unwilling to give them their first professional opportunity. This is a mistake and a missed opportunity to capitalize on this new source of software engineering talent.</p><p>If you look at any job in any field at any level of management, the same set of traits are common to every high performer. High performers are self-starters, hungry for more challenges. They take risks, they have grit, and they learn quickly. Career transitioners are flush with these traits. We know because of the journey they’ve been on, the decision they made to pursue software engineering, and their success in the intensive coding curriculum. They will be more than solid contributors — they are the future leaders of our engineering teams.</p><p>In addition to having the right core traits for success, career transitioners bring valuable skills and perspectives with them to their new software engineering jobs. There are generic skills like collaborating and communicating effectively with teammates, empathizing with the customer, and analyzing data. These skills are critical to being an effective software engineer, and many career transitioners are already experts. Additionally, there is often a specific alignment in prior experience and needed expertise. A former accountant’s perspective and industry knowledge will be particularly valuable at a fin-tech startup. The same is true for a former nurse at a health tech company or a former sales rep at a CRM company.</p><p>If for no other reason, the software industry should invest in hiring career transitioners because they’re available and ready to start today. Some forward-thinking companies have set up apprenticeship programs that are designed to attract talent from this rapidly growing source, but this is not yet the norm. Hiring many great engineers is one of the hardest problems that engineering leaders face. It’s time to take a different approach.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b1b02dc7a51b" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>