Acerca de
As a Lead Engineer, I combine a solid…
Actividad
240 seguidores
-
Mikhail Diakov ha publicado estofun experience with code review from Copilot Copilot's review notes: The validation loop could be optimized by combining all validation calls or implementing early exit. If any product fails validation, the subsequent validations are unnecessary since the process will be skipped anyway. _______________________________________ My review notes: Your code changes caused the violation of database transaction borders. Before your changes it was: ``` BEGIN TRANSACTION SELECT ... subscription FOR UPDATE UPDATE ... subscription ... COMMIT ``` after you changes it's now: ``` BEGIN TRANSACTION SELECT ... subscription FOR UPDATE COMMIT BEGIN TRANSACTION UPDATE ... subscription ... COMMIT ``` please change it back
-
Mikhail Diakov ha compartido estoThe great talk with Andrej Karpathy about what's the difference between human and what we call AI, why reinformcement learning is stupid and evolution of intelligence A few quotes I liked: We're (humans) not actually that good at memorization, which is actually a feature. Because we're not actually that good at memorization we're forced to find patterns in a more general sense Human maintain a huge amount of entropy You can't just keep scaling "reflection" on the same amount of prompt information https://lnkd.in/dKt-smFdAndrej Karpathy — “We’re summoning ghosts, not building animals”Andrej Karpathy — “We’re summoning ghosts, not building animals”
-
Mikhail Diakov ha compartido estoMikhail Diakov ha compartido estoI don't normally post on LinkedIn but came across an article that really resonated with me about some key things that make the difference in a senior software engineer. It is by no means a complete list, and an updated note from the author at the end of the article even recognizes this, but still a worthwhile read. https://lnkd.in/gYC9VFCg
-
Mikhail Diakov ha compartido estoSpeaking about tech debt (and therefore the increase of costs for development), one of the major sources of tech debt is unclear and\or weakly specified team boundaries. Unclear boundaries appear when each team doesn’t know which part of the system is their responsibility and which belongs to other teams. This can lead to several negative consequences, such as: - Parallel, conflicting changes to the same module or class by different teams, resulting in merge conflicts, regressions, or broken environments. - Stepping on each other’s roadmap, where one team unintentionally blocks or undoes the work of another. - No clear ownership for bugs or incidents, which slows down investigation and resolution. - Cross-team dependencies that are discovered too late, leading to rework and delays. - Critical features or services degrade over time due to a lack of explicit responsibility. Weakly specified boundaries make it unclear how to keep different parts of the system—owned by different teams—isolated from one another. As a result: - Teams introduce changes that create tight coupling between modules they don’t own. - Interfaces or contracts between components evolve inconsistently - Shared code becomes a dumping ground where everyone contributes, but no team takes accountability. - Refactoring becomes risky and slow, because no one knows who should approve or implement the change. - Accumulation of ad-hoc fixes instead of cohesive design decisions, increasing long-term complexity. Ultimately, unclear and weak team boundaries increase coordination costs, slow down delivery, and generate long-term architectural and organizational debt. Here is a good talk about this topic https://lnkd.in/dpmYamtPIntroduction to Context Mapping - Michael Plöd - DDD Europe 2022Introduction to Context Mapping - Michael Plöd - DDD Europe 2022
-
Mikhail Diakov ha publicado estoWhile there is a race for AI adoption to increase the business margin and reduce expenses no one has canceled the tech debt for software companies and its impact on margin. And even vice versa vibe coding potentially is huge source of tech debt in exchange of rapid prototyping and development McKinsey surveys showed that CIOs estimate that technical debt accounts for 20–40% of the value of their technology assets, and it continues to grow in most organizations. (see https://lnkd.in/dS68ci89) So the question how to manage the tech debt is still actual. Here is relatively old (and I believe not so popular as AI articles) but still intresting article from Google regarding their efforts to manage tech debt : https://lnkd.in/danbEA5Q (free to read) Also I recommend the interesting study "Business-Driven Technical Debt Prioritization", it's paid and can be found here for example https://lnkd.in/d8NYgD-i
-
Mikhail Diakov ha compartido estoMikhail Diakov ha compartido estoTo vibe or not to vibe? I wrote a new memo about the constant little risk assessments I make during AI-assisted coding, thinking about probability and impact if AI gets it wrong, and if I will be able to detect that. https://lnkd.in/d_JhrWhX
-
Mikhail Diakov ha compartido estoA few words about why code quality is important for the business, among other things https://lnkd.in/d6k5isBy
-
Mikhail Diakov ha compartido estoMeasuring the impact of AI for software development https://lnkd.in/d4dsSyDzMeasuring the Impact of Early-2025 AI on Experienced Open-Source Developer ProductivityMeasuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity
-
Mikhail Diakov ha recomendado estoMikhail Diakov ha recomendado esto⚡ 9 teams. 7 hours. 1 Hackday. This week, we brought the company together for a dedicated hackathon aimed at tackling one of consumer finance’s most pervasive challenges. Enabled by AI, in just a few hours, each team was able to build, test and pitch working prototypes to our panel of judges. Sharp thinking. Big ideas. And real-world impact. Huge credit to everyone who stepped up 👏 #Fintech #AI #Innovation #Wealthtech #FinancialAdvice
-
Mikhail Diakov ha recomendado estoMikhail Diakov ha recomendado estoTech AI innovation — that delivers value to customers. With SmartAccept, the team at Unbiased has developed an AI feature that operates quietly in the background to help advisers secure more of the right clients. It’s a great example of how machine learning can be applied with precision: learning adviser preferences, automating accepting enquiries out of hours, SmartAccept helps firms grow efficiently without compromising on quality or control. This is what meaningful innovation looks like in financial advice. Huge kudos to the team for using the latest Google Cloud Platform AI technologies (Vertex AI, Gemini Embeddings), to deliver a big customer experience win. SmartAccept is currently available in beta, on an invitation-only basis to UK-based financial advisers and mortgage brokers. #SmartAccept #AIinFinance #ClientGrowth #GCP #AI #VertexAI #GoogleGemini
-
Mikhail Diakov ha recomendado estoMikhail Diakov ha recomendado estoAI I have been in software development for almost 60 years. Most of my friends have been in the profession or in academia. It is not at all surprising that AI has been central to a lot of conversations. I do not really want to engage in an extended debate—we will not need to wait very long for history to provide some semblance of answers—but do want to offer one point of view. My Masters degree in Computer Science was in AI: toward the end of generation one (vision, speech, neural networks and simpler machine learning) and the beginning of generation two (neural networks, more interesting machine learning, and "Expert Systems"). Expert Systems like Mycin routinely outperformed human diagnosticians. Business and stock trading Expert Systems were touted as superior to human managers and traders. Advocates of the technology boldly predicted that any company not employing Expert Systems would be out of business within the decade (this was the late 1980s). The hyperbole then was as loud as with today's carnival barking for LLMs, but not quite as widespread. Obviously, the myriad failures and unrealized promises let to a deep AI Winter. I was a skeptic and critic then and remain one today. My first professional publication (1989, a two part article in AI Magazine—then the journal of record for the field) offered a position that I still believe relevant today. When Alan Newell entered his classroom and told his students, "over break, Herb Simon and I created thinking computer,"—referring to their general problem solver—he was making the mistake of believing that the way he thought, scientifically, logically, mathematically, was the epitome of human thinking. Today's touters of AI make a nearly identical mistake. Equating the manipulation of symbols (mostly language) and algorithmic thinking is, not only the supreme example of human thought, but the only one of any importance. In my opinion, AGI will be achieved only when we have demeaned and debased human intelligence to the degree that a machine might be able to emulate whats left.
Experiencia y educación
-
Unbiased
Mira la experiencia completa de Mikhail
Mira su cargo, antigüedad y más
o
Publicaciones
Idiomas
-
Spanish
Competencia básica
-
English
Competencia básica profesional
-
Russian
Competencia bilingüe o nativa
Recomendaciones recibidas
2 personas han recomendado a Mikhail
Unirse para verloPerfiles similares
Ver más publicaciones
-
Vistatec
Software teams are under constant pressure to move faster, and translation often begins after UI decisions have been made and strings have been exported. This results in delays, rework, and avoidable quality issues. https://lnkd.in/dwKaZn4s Vistatec and Lingoport’s new partnership brings shift-left localization into the heart of product design and development by combining the power of Lingoport Localyzer with VistatecVerifier. #Figma #Localyzer #VistatecVerifier
-
IT DAYS
Vlad Mihalcea is a Java Champion, creator of Hypersistence Optimizer and Hypersistence Utils Open-Source project. He is passionate about communities, technical leadership, system architecture, and database systems. He wrote hundreds of articles about Java, JPA, Hibernate, and SQL on his blog, and he is the author of the High-Performance Java Persistence book and video course series.
-
Enigma Security
🚀 Accelerating Spring Boot Applications: From Minutes to Seconds with GraalVM In the world of Java development, performance is key for scalable applications. Recently, we explored how a team optimized a Spring Boot app, achieving an impressive acceleration of up to 10 times using GraalVM. This native tool compiles Java code into efficient binaries, reducing startup time and resource consumption. 🔍 The Initial Challenge - Traditional Spring Boot applications take minutes to start due to dynamic class and dependency loading. - In cloud environments or microservices, this generates latency and unnecessary costs in CPU and memory. ⚡ The Solution with GraalVM - GraalVM enables ahead-of-time (AOT) compilation, converting the code into native executables that start in seconds. - Integration with Spring Boot Native: Supports native profiles for automated builds, maintaining compatibility with existing frameworks. - Measured Results: 90% reduction in startup time and up to 50% in memory footprint, ideal for Docker containers. 📊 Practical Benefits - Greater efficiency in CI/CD deployments and horizontal scalability. - Lower environmental footprint by optimizing cloud resources. - Facilitates migrations to serverless architectures without sacrificing functionality. This approach not only accelerates development but redefines performance standards in modern Java. For more information, visit: https://enigmasecurity.cl #SpringBoot #GraalVM #JavaPerformance #SoftwareDevelopment #Optimization If this content inspires you, consider donating to the Enigma Security community to continue supporting with more technical news: https://lnkd.in/evtXjJTA Connect with me on LinkedIn to discuss trends in cybersecurity and development: https://lnkd.in/ex7ST38j 📅 Sun, 11 Jan 2026 08:30:31 GMT 🔗Subscribe to the Membership: https://lnkd.in/eh_rNRyt