New TIL: Positioning is a common problem for data scientists

content/til/2023-12-18-positioning-is-a-common-problem-for-data-scientists.md

@@ -0,0 +1,33 @@
+---
+title: Positioning is a common problem for data scientists
+author: Yanir Seroussi
+type: til
+date: 2023-12-18T00:30:00+00:00
+url: /til/2023/12/18/positioning-is-a-common-problem-for-data-scientists/
+summary: "With the commodification of data scientists, the problem of positioning has become more common: My takeaways from Genevieve Hayes interviewing Jonathan Stark." 
+showBreadcrumbs: true
+tags:
+  - business
+  - career
+  - data business
+  - data science
+---
+
+I [became a data scientist by accident](https://yanirseroussi.com/2015/05/02/first-steps-in-data-science-author-aware-sentiment-analysis/): I followed my curiosity and did a PhD in computational linguistics and recommender systems. When I finished my PhD in 2012, I discovered I could call myself a data scientist rather than _a software engineer with a research background_ (which was a bit of a mouthful). As 2012 was the year [Harvard Business Review declared data scientist to be the sexiest job of the 21st century](https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century), I didn't need to think much about what kind of data scientist I was. Just _being_ a data scientist was pretty unique.
+
+The world has changed in the past eleven years, and now there are many more data scientists. While you could earn good money as a generic data scientist, you don't stand out. That is, it's not only that [software commodities are replacing interesting data science work](https://yanirseroussi.com/2020/01/11/software-commodities-are-eating-interesting-data-science-work/), and that [large language models are making some skills irrelevant](https://yanirseroussi.com/2023/04/21/remaining-relevant-as-a-small-language-model/) – whatever is left of the core data science skillset has become an undifferentiated commodity. 
+
+I've been thinking a lot about positioning as an independent consultant recently, after realising that [the lines between solo consulting and product building are blurry](https://yanirseroussi.com/til/2023/09/25/the-lines-between-solo-consulting-and-product-building-are-blurry/). One great source to learn more on the topic is [Jonathan Stark](https://jonathanstark.com/), who has published many valuable resources over the years. Among them, I found a podcast interview he did in May this year with Genevieve Hayes, titled [Building Your Authority in Data Science](https://www.genevievehayes.com/podcast/ep14/).
+
+Whether you're an employee or independent data scientist, it's worth listening to the interview. Here are my key takeaways:
+- Even though data scientists are already highly specialised, the problem of positioning oneself and standing out is common.
+- Understanding marketing and the business side in addition to mastering the technical skills can be a superpower, as you can act as a bridge between non-technical people and the "nerds".
+- You need to be perceived as meaningfully different _by your target audience_, regardless of whether you choose to specialise in a horizontal (specific data science skill like computer vision) or in a vertical (specific industry like renewable energy). If your target audience doesn't find you meaningfully different, you have more work to do.
+- Avoid basing your self-worth on where you sit compared to other data scientists. If you're good enough technically (C to B+) and you have complementary skills and an outcome-driven mindset, you'd be unstoppable. This still seems rare.
+- Publish stuff that business people can understand, i.e., connect what you can do on the technical side with business value.
+- You don't need to be managing people to deliver results, e.g., Jonathan chose to remain solo and not hire employees. Focusing on business results is what matters.
+- At the time of the interview (May 2023), Jonathan was searching for a ChatGPT consultant to learn whether he could turn his content into a chatbot. He was surprised he could barely find anyone. This is a good example of riding a hype cycle, as being an early authority on ChatGPT can lead to solid business outcomes for indie consultants. However, given the nature of hype cycles, this can change quickly.
+- Do things you're deeply curious about, as enthusiasm helps you stand out. Going down rabbit holes can be a strength. In the context of ChatGPT consulting, this reminded me of [Simon Willison](https://simonwillison.net/) and [Ethan Mollick](https://www.oneusefulthing.org/), who have become more well-known recently due to their curiosity and blogging on generative AI.
+- There are many ways to make money, so you might as well work on something you like.
+
+Personally, I'm still figuring out my positioning. I found it interesting that Genevieve and Jonathan agreed that stereotypical data scientists care more about building their models than about how they're used. While I enjoy the technical aspects of modelling and other data science tasks, I'm much more interested in shipping work that matters. That's why I became a data scientist originally – I left academia after my PhD and joined startups to get close to business problems and build stuff people use. If that's still a rarity, I suppose it can help with my positioning. That said, I'm also exploring a deeper vertical specialisation (currently looking at energy markets).