Michelle Rascati

In our recent team retrospective at DNSimple, we recognized a growing concern: technical debt. Many teams face this challenge, but rather than viewing it as a burden, I approach it as a strategic deferred investment in our codebase's design. In my latest blog post, I define technical debt, its impact on software projects, and how we can align its management with our product and engineering roadmap. This post is for developers and non-technical stakeholders who want to understand the nuances of technical debt and how to manage it effectively. Check it out, and let’s continue the conversation on innovating sustainably while managing our technical debt wisely. https://lnkd.in/dcihZ4fE

8

Whenever I get "worried" that an LLM will take my job as a SWE in the near future I look back at this benchmark leader board of real-world issues and am reminded that LLMs have been stuck at making marginal improvements bellow 20% resolved rate for over a year now. That being said, I do think AI will become an increasingly powerful tool in a SWEs belt in the coming years if it hasn't become one already. I personally use it to help me explore a problem space faster and come up with boilerplate for trivial problems and expand from there. In those cases it's like using Google and Stack Overflow and other forums, but faster. Edit: I'd like to also point out that I really like this benchmark because you cannot train a model on it unlike HumanEval. It's in essence, a true test of an llms ability to generalize and solve problems. Source: https://lnkd.in/epK2GQ2w

2

I started writing regularly about AI. If you're an innovators or product developer eager to apply the latest in AI, whether you're in a startup or a big company, consider checking https://lnkd.in/gGdW8PYe I won't teach you ML, but I share real-world insights from launching - and sometimes failing at - numerous ML features on a large scale.

55

2 Comments

Really excited to share this article which Zhou (Jane) Xing, Andy Woods, and I wrote which highlights the data movement tool we've been working on -- MOLT Fetch. It's exciting to build tooling that simplifies our customers' migrations over to CockroachDB. In this post we discuss how our data movement story has evolved and how Fetch aims to solve issues we discovered along that process. If you're interested, you can read more here: https://lnkd.in/gCzDKU8G I'm also curious if anyone in my network has experienced challenges with performing data movement and migrations, and what some of the biggest gripes have been?

21

When reading the manual doesn't help. It's NOT MY FAULT this took longer than it should have. I read the quick start, but not the actual pages. It should be enough. But now you, too, can save time getting blob counts from Azure with go :). https://lnkd.in/g4TzPitK

2

I relate to this quite a lot. I'm currently migrating some core logging functions, for example, to fix some systemic issues, and while I'm not doing all of the labor myself, I was certainly the only person driving the change, implemented the new methodology, and am handling all of the issues which arise in the process. Most of my projects these days (often self-directed) are of this type (ie: see something which needs systemic improvement, even if it touches a lot of code, and figure out how to do it).

2

When I first learned about Amplitude, I was an engineer at another company. We were told we had to integrate this new analytics tool, and the reaction across the team was, "Where did this come from, and why do we have to do all this extra work?" Fast forward a few months, and we were all hooked—engineers, product managers, everyone. The effort spent integrating Amplitude was absolutely worth it. Suddenly, we could easily analyze our data, and it was miles ahead of anything we'd used before. But here's the thing—many Amplitude customers struggle to get past that first hurdle of implementation. So, we made it simple. Now, you can implement Amplitude with just ONE line of code! And don’t worry, precision tracking is still there—we’re just making it faster and easier to reach the moment where you can start analyzing your data. Learn more about #AmplitudeMadeEasy: https://bit.ly/3AUKC37. 

32

3 Comments

Its about solving the problem, not 'having AI' in the system. This is from last March, but the Wave 1 vs. Wave 2 distinction does seem to be playing out. Less 'generate me content', more 'synthesize vast amounts of relevant data and aid in decisonmaking or signal extraction'. https://lnkd.in/ehFBFC32 They don't talk about multi-agent systems here, but looking forward to more startups with a focus on leveraging agentic design patterns (https://lnkd.in/ed5rMmFe) to streamline real world workflows.

4

1 Comment

I just finished reading System Design Interview – An Insider's Guide: Volume 2, and I can’t recommend it enough for anyone looking to improve their technical problem-solving skills. It is an excellent read for those looking to excel in system design. It’s packed with useful insights and practical advice that can help you tackle technical problems more effectively. Happy reading!!!

16

I have a strong opinion about this. I have been a senior engineer for most of my career and am always looking for feedback to achieve a promotion to principal engineer. I agree that impact is more important than output, but it's hard without being a principal engineer. If you are not a principal engineer, you won't have access to meetings and knowledge to distinguish what is impactful and what is not. Until then, you need to have a mentor, a manager, or someone with a bigger picture who is able and willing to guide you in discerning what is impactful and what is not. If this doesn't happen, your only option is to increase your output, hoping that one of those things will have a considerable impact by chance.

13

This is a fun blog post about first steps with Rust -- and what looks like a legitimately useful tool!

22

TL;DR Impacted by the Tesla layoffs. A wild ride worth reflecting on and taking some time to recharge the batteries. It all started innocently enough with a LI message from a Tesla recruiter in the early months of 2020, and little did I know the entire world would be turned upside down for the better part of 2 years. What was supposed to be an onsite interview suddenly became a virtual interview, and ended up as a WFH onboarding. Tasked with building a scalable Learning platform from scratch for our almost 160k+ strong global workforce of employees and partners as we migrated systems with an early 2021 delivery deadline seemed an almost insurmountable ask at the time, very much complicated by being a new hire and trying to get things done in a virtual work environment the entire company was figuring out. Against this backdrop, I got to witness first-hand the undeniable dedication, efficiency, and passion of all of my colleagues at Tesla. There was no hesitation when we had to spend late nights until 3am during Thanksgiving week nor the countless weekends collaborating to ensure we got the job done. And got the job done we did. Fast forward 4 years, and together we’ve built an enterprise Training platform that has scaled to support Tesla’s almost 300k global workforce of employees and partners through eLearning, In-Person instruction, and on-the-job Assessments. We had the unique opportunity of supporting all parts of our business globally from HR to Engineering to EHS&S to Energy to Manufacturing to SS&D and to Supply Chain. While it’s certainly a disappointment for myself as well as for my many dedicated and passionate colleagues who have been impacted, I take immense pride in knowing that we all contributed to a greater cause and mission than ourselves: Accelerating the world’s transition to sustainable energy. And even though we’ll no longer carry the torch for now, I know that through our many contributions that we’ve left Tesla in a better place than when we joined, and Tesla is absolutely primed for the next growth phase. We’ll all be cheering from the sidelines! I would be remiss if I did not mention some of the talented colleagues who were instrumental to my Tesla journey: Malia S. Eben Carey Dave Parrent Shelby Callahan Avinash Matha Janelle Flores Lixin Wang Prateek Khanna Pankaj Saini Ghizlane Bounoua Aarthi Palaniswamy Renee Hu Lizette B., PsyD Molly Caddam David Dodge, PMP Dillon Lee Matt Primeau Jack Sylvester Kevin McMurphy Alfredo Aguirre Stephanie Hernandez Divya H Manasvi Komma Mike Nobles Carly Bermensolo Rafae Shaik Xiechuan Liu #Tesla #Teslalayoffs #TSLA

477

37 Comments

I'm fast approaching my #opensource sweet 16. 16 years doing work under permissive license, where everyone can see your faults. To me, getting started the biggest hurdle was that I could do anything. People are so much better than me at everything. I feel like an idiot (because I put myself lower than others). I just am happy someone responds. When they ask me to create code, I feel like the war boy in mad max, like a god just looked at me, shiny and chromed. Then, reality. These are real humans, often borrowing time. You are actually helping, but also codependent. Maybe you feel you aren't pulling you weight, there's still an imposter syndrome but you know with enough work you can at least "cover your costs". You don't pay much attention to the time spent, maybe you should have. Then, leading. Wow people are great and weird and difficult. Sometimes your entire day has nothing to do with lines of code, and you try to figure out how to make your way through it all. Perhaps like me, you optimize for users or folks who are more reliable. You make choices, none are great, but the energy is rewarding as are the accomplishments. You feel a higher sense of duty and maybe debt. Then, community. No project is an island, and users are not sufficient to achieve a goal. You also recognize not everyone cares about your project, but maybe they care about open source in general, or the language, or just friendship, it doesn't matter. You still do the work on code, you must, and all the drama, but you also spend time to make sure what you do has impact. Also you want advice from other people through all these things. Maybe, burnout happens next ;) In my case, several times or stopping or whatever. You have the shame you feel, like you should be superhuman and just take whatever because you started something. Life happens, you realize there are constraints, choices, there's no easy decision. If you feel up to it you get a second, third or fourth wind. You keep your optimism despite the reasons to be careful. You try again, re-engage, give it your all, but not 100pct of your all. You learned that doesn't work, so you try to learn what can work. You still want the ideals to occur but ack life is complicated, for you and everyone else. You do your best, you try to keep real while incentivizing folks into a start you had 16 years ago, or a maturity you saw in the past or wish you had. That's where I'm at! I hope the story is interesting

29

I listened to an interesting podcast this morning about the emerging job title of "AI Engineer". This podcast resonated with me and I thought I should summarize some of my thoughts on the topic here. The Evolution of Tech Roles: From Webmaster to AI Engineer Job titles are constantly evolving. This evolution brings an important realization: as roles specialize, the perceived entry barriers become lower. The 90s Webmaster: In the early days of the Internet, there was one person who handled everything web-related and had the title "Webmaster". This magical person was expected to build, maintain, market, and optimize the company website. This role has since fragmented into specialized positions like front-end developer, back-end developer, UX designer, and various digital marketing specializations. The 2010s Data Scientist: In the early 2010s, a data scientist was described as someone who would do data analysis, realize a business opportunity, build a machine learning (ML) prototype, communicate the opportunity to leadership, then productionize their system and manage the new ML system. This role has now split into various positions such as data analyst, data scientist (focused more on statistics, visualization, and business communication), data engineer, machine learning (ML) engineer, and ML researcher. An ML researcher, Boris Meinardus, made a great video on these roles for those who want a more detailed take: https://lnkd.in/eGCRf5MW. The 2020s AI/ML Engineer: We're now witnessing a separation between the ML engineer, who builds and maintains more traditional ML systems requiring a deep understanding of mathematics, statistics, and ML algorithms, and the AI (LLM) engineer, who focuses on LLM-powered application development. To become an AI (LLM) engineer, you need to learn to use LLMs and their associated tools and paradigms: retrieval augmented generation (RAG), LLM tool chains (like LangChain), and associated technologies like vector databases. Consider the database analogy: Most developers use databases without fully understanding their inner workings or ever implementing B-trees. There are database experts, but for most developers, databases are useful black boxes. Similarly, to start in AI engineering, you don't need to unravel every layer of ML complexity. It's great if you have that knowledge, but it's not a prerequisite for building useful LLM applications. This shift opens doors for more developers to enter the AI field. Obviously if you want to develop the core LLMs and take an ML engineer job on a team in a company that builds foundational models (OpenAI, Anthropic, Google, Meta, etc.) then you do need the fine grained, detailed understanding of the underlying math and algorithmic inner workings. This would be the same if you wanted to do cutting edge database R&D. There is also another interesting post on this topic here: https://lnkd.in/eyPyaQ9S Thanks to: James Brady Adam Wiggins

102

6 Comments

Have you checked out the latest posts on this GitHub repo? They're quite impressive. It reminds me of my time at S3, where durability reviews were a top priority. Every developer at S3 had a deep sense of paranoia about durability and was always looking for ways to strengthen it. That mindset is ingrained in us and will never go away. It's great to see that same level of dedication in the open-source community. #opensource #S3 #durabilityreviews

12

I am back to help fellow techies with mock interviews! If you are a senior engineer, tech lead or an engineering manager currently pursuing opportunities and interviewing with companies? I can help by mock interviewing you (no cost). Topics are: - Engineering Management - System Design - Coding I will share feedback, what went well and areas of improvements. If you are not currently preparing, please share with your friends or folks who might need this. Comment for reach, or share in your network. Book a 45/60 mins time slot with me. (link in comment). #softwareengineering #softwareengineers #interviews

87

13 Comments

An insightful perspective on developing a "taste" for product design. This could apply to crafting clean, elegant code or shaping how users interact with your product. Why not experiment in your next hackathon or work on a personal project to refine existing tools—internal or external? Think about what kind of standout experience would captivate users, something surprising and delightful enough that they’d spread the word and share it with their team. How can you create something truly memorable? https://lnkd.in/eD8pStVr

1