Karachi Wala Developer

As AI agents become capable of rewriting entire C compilers and Next.js repositories in days, the value of software is shifting away from the code itself. If LLMs write probabilistic code, how do you guarantee deterministic results? The answer: Tests. In this episode, we argue that developers need to stop agonizing over AI-generated implementation code and instead double down on writing, reviewing, and curating rigorous test suites. Learn why some open-source projects are actually closing their test repos, and how adopting an Architect mindset will unlock true AI autonomy in your workflow.

The role of the developer is changing. In this episode, we discuss how to stop being a "consumer" of AI prompts and start becoming a "producer" of AI systems. By shifting your focus from writing code to designing autonomous loops and persistent architectures, you can move up the ladder of value and start steering the future of software development.

As software engineers, we are trained to control every detail—from code formatting to system architecture. But autonomous AI demands the exact opposite. In this episode, we discuss the "Autonomy Paradox." Learn why enforcing arbitrary rules on AI agents causes them to fail, and how shifting your focus from "how it's built" to "what it achieves" is the key to unlocking next-generation development.

If you are dumping your entire codebase into Claude or Gemini and hoping for the best, you are doing it wrong.

In today's episode, we tackle Context Engineering. AI agents have short-term memory (the Goldfish Problem) and limited Context Windows. If you fill that window with unnecessary data or heavy MCPs, the AI's performance drops drastically.

Tune in to learn how to structure your information using Agent files, Skill files, and Rule files. We also break down advanced strategies like Progressive Disclosure (only giving the AI what it needs, when it needs it) and using Sub-Agents to keep your main context window clean.

In this episode, I share a personal story of building an ERP system with AI in the field, and how it forced me to confront the "5 Stages of Grief" as a veteran software developer. The way we have written code for the last 20 years is ending. To survive, we have to unlearn our "deterministic" mindsets and adopt new mental models for a non-deterministic world. Join me as we kick off a brand new series on the foundational literacy required for the future of tech.

Dive into the much-anticipated GPT-5 update. We cut through the hype to analyze OpenAI's latest offering, covering its new unified architecture, efforts to reduce hallucinations, enhanced memory, and the controversial decision to deprecate GPT-4. Is GPT-5 the leap towards AGI we've been promised, or just an incremental step? We compare its capabilities against Gemini, Grok, and Claude, and share our candid thoughts on whether it truly delivers or leaves us more excited for upcoming models like Gemini 3.0.

Dive into the AI coding revolution! This episode explores how AI is transforming software development, from rapid prototyping and coding assistance with tools like GitHub Copilot, to the rise of generative coding. Learn the key trends, essential tools, and practical ways you can leverage AI to build faster and smarter. Perfect for developers and anyone curious about the future of code.

Ever wished your AI could actually do things directly from your terminal, beyond just chatting? Google's new Gemini CLI is here, and it's a game-changer! Unlike previous closed-source tools, this open-source power tool brings the intelligence of Gemini Pro right to your command line. We're talking direct file system access, automated project setups, and a powerful "reason and react" loop that lets AI analyze, plan, and execute tasks on your machine. Perfect for DevOps, developers, and anyone ready to automate their workflow. Is your terminal ready for its new brain?

Are Large Language Models (LLMs) truly intelligent, or just sophisticated pattern matchers? This episode dives deep into a fascinating debate sparked by Apple's recent research paper, which questioned the reasoning capabilities of LLMs. We explore the counter-arguments presented by OpenAI and Anthropic, dissecting the methodologies and the core disagreements about what constitutes genuine intelligence in AI. Join us as we unpack the nuances of LLM evaluation and challenge common perceptions about AI's current limitations.

Dive deep into the world of AI communication with this episode, focusing on the game-changing MCP (Model Context Protocol). Discover how this protocol, pioneered by Anthropic, acts as a universal translator, enabling AI models to seamlessly interact with external services like Google, Twitter, or even your private databases. Learn why MCP was developed to overcome the limitations of static API documentation, ensuring consistent and accurate information for AI. We'll explore its client-server architecture, how it empowers custom integrations, and why it's becoming the core of connecting LLMs to the vast digital world. Get ready to understand the future of AI interaction!

Dive into the rapidly evolving world of Large Language Models! This episode breaks down the latest updates on cutting-edge models like Gemini 2.5 Pro (with its 2 million context window), the ambitious Llama 4, GPT-4o's impressive image generation, and Cloud's Sonnet 3.7. We explore key trends such as expert architectures, the nuances of leaderboard performance (including Llama 4's controversy), the emergence of "deep thinking" AI capable of independent research, and the potential impact of upcoming models like GPT-5. Discover how these advancements are shaping the future of AI.

In this episode - I take on the task of explaining what vibe coding is, and if we should keep studing software engineering. I introduce some vibe coding tools, including the latest one Firebase Studio from Google

Last week OpenAI released a new model called O1 - which has advanced capabilities to reason and does a lot better in maths and coding. How does it do this and what does this change? This and more in this episode.

In the world of many open source LLM models, Google Gemma stands out with some unique capabilities. Firstly its made by Google after all - so its implementation is fully optimized and memory efficient. This and more on what is Gemma about and how its better.

This episode dives into the exciting world of on-device AI, specifically Large Language Models (LLMs) running directly on your smartphone! We'll explore why cramming this powerful tech into your pocket is a game changer, and discuss how LLMs can transform your daily interactions with your phone.

In this episode we cover the latest updates from Google IO and OpenAI Spring update. The all new ChatGPT-o and Gemini 1.5 Flash. We talk about where are LLMs headed next.

Lama3 is a popular open source model by Meta, the company behind Facebook, Instagram and Whatsapp. They have release the next iteration of the famous Lama2 model, this is a popular open source LLM which is heavily used by developers world wide. Is the new iteration strong enough to take on the best closed source models? We explore in this episode.

As we move to use LLMs more and more for different use cases, we will run into at times issues or limitations with the model. There are many different ways to train the model to do a specific task better, and fine tuning is one of them. What is about, when do we need to use it and how do we do it - all in this episode.

This episode covers my journey as a GDE over the last few years, what got me into it, and what is the big outcome. Also joined another category for GDE in Machine Learning.

In this episode I share my thoughts on the latest additional to AI development called Deven, and where the industry is moving towards. Will software engineering be taken over by AI in the next few years?

So this week we had 2 back to back announcements which completely shook up the LLM worlds. Gemini Pro 1.5 with its 10 million token context length makes it super powerful to run on large codebases and a lots of documents. The other big news was the launch of OpenAI Sora - their new text to video generation which can now generate super realistic videos for up to a minute. How will these impact our tool, I share some thoughts here.

The all new heavy weight model has just dropped this week. Should you be rushing to try it out? Why is the ultra model better? This and more in his quick episode talking about Gemini Ultra.

In this episode - we discover the Retrieval Augmented Generation (RAG) process which allows us to train our LLMs using our data. In this short episode we cover what is this about and how can we use it.

If you are using ChatGPT or Bard - you are "prompting" the AI to get some answers. Not all prompts are made equal. The response of the LLM depends heavily on the quality of the prompt and the context given to it. Thus its worth while spending some time understanding prompt engineering and how to do it properly.

Referencing the classic episode from IT Crowd - we take our first dive into the world of LLMs by talking about neural nets and how they form the basis of the LLMs we know today. We touch on the famous paper - All you need is attention, and then talk about what LLMs under the hood are really about.

After a year long break - Im back with an update on Gemini and Gen AI taking over. In this episode I talk about the over all impact of Gemini and LLMs and how can we use them better in our daily lives.

I had a great chat with Lamak, who works at one of the startups in Karachi called Bazaar Tech. We discussed several interesting topics including team structures, full stack development, performance reviews and more! Thanks Lamak for taking time out for this episode of EngineeringLeadershipPakistan.

I had a great chat with Asadullah about his time in Airlift where he was an engineering manager leading a specific team. We talked about several management related topics, from the values at Airlift, life there as an engineering leader,  performance management, hiring, engineering productivity and more! 

You can also watch this on the KarachiWalaDeveloper Youtube channel.

This is the first episode of the all new series where I will be interviewing Engineering Leaders in Pakistan. We will be talking about leadership topics like CTO roles, hiring, performance reviews, feedback and more. In this episode we have Hunain Kapadia from Smart Benefits sharing about this experience building the some awesome tools in the insurance space.

Software Engineers require instructions in order to build apps and tools. And most of us don't know what these instructions should look like. One of the ways these instructions can be shared is in the form of user stories, and it's not just about writing a bunch of text, but comprehensive check list of how this should be structured so the right information can be given to the builder. Ofcourse its a two way effort Product and Engineering need to work closely to build up such requirements so they make sense from a functionality perspective, and a technical perspective.

This session is based on the article linked below:

https://martinfowler.com/articles/product-backlog-building-canvas.html

Google, or any of the FAANG companies for that matter, are an awesome place to work. When I started by bachelors - I was thinking of graduating and then getting a job there, or build a pathway to a job there. However by the time I graduated, my mind set had changed. In this episode I explore my thoughts and reasoning on what changed my mind and why I havnt applied since.

With startups going bust left right and center, high inflation and a global recession - are startups still the right choice to be at? In this episode I talk about the difference between the companies treating engineering teams as a profit center vs a cost center, to show why startup engineering teams as usually a better choice even if there is a slightly higher risk being there.

In this episode I break down a good engineering culture using two different culture tests from the web. We take a deep dive into the technical side of the culture first, followed by the work and management culture. How does your company score - what is the one thing you would change in your company to make it a better experience?

Javascript is one of my favourite languages out there. Im not sure if it just the familiarity with the language (have been using it since age 10) or the fast pace of change in the language space or the fact its being used on space craft - I think its a useful skill to have for anyone working in the software industry. Check out my rant on what makes JS so awesome according to me!

As programmers we work hard all day to ensure we build the best products in the best technical way possible, and if that work isn't shipped to customers right away, we are unable to generate any value of it for the company / product. DevOps plays a critical role in our organisations today, and ensure we have an efficient production line shipping features to our customers is a part of it. In this episode we review how we can better understand and measure performance of our software delivery and what does it take to compete with the best out there.

While mentoring some students with technical interviews, I documented what I find to be most important qualities of a software developer or engineer. While a quick google search online will find mixed reviews, most people feel that technical skills are most important. However I feel a bit differently about this, and in this episode share the things all software engineers should focus on and how they can improve them selves if lacking.

The recent failure of the startup Fast got me thinking of things one needs to be aware of while considering to join a startup and things to observe while working in one. This episode covers some basics to what you can do before and after joining which should help you prepare better for the risks involved while working in a startup.

This episode was inspired by the following post: https://newsletter.pragmaticengineer.com/p/the-scoop-fast?s=r

Technical Due Diligence (DD) for startups is not something we think about until the time really comes in. When we get the list of questions or get asked those dreaded questions that we always knew would come to haunt us. In this episode I share about what is the process and what to expect from this, so that one can be better prepared when this happens.

T-Shaped developers are those interested in learning across a breadth of technologies ranging from frontend to backend to devops. In this episode we try to understand why would we want to diversify vs specialise and is there a better option between the two?

Do you feel your team is very sluggish and not moving fast enough? Do you feel there are constant blockers which are stopping the features to get shipped? In this episode I dive into the world of engineering performance and what are some thing of the useful things you can measure, and which ones are not meant to be measured at all.

You might have been reading about one of the dozen multi-million dollar investments made into startup recently. And you might be wondering how do I get a piece of that action. Most companies now offer their shares to employees through whats called an ESOP. This allows you to accrue value over time and if the company wins big - so do you! This and more in this episode.

PCI DSS is a security standard certification which a company requires if they have card payment integrations on their websites. However its one of those topics which sounds scary, not talked about enough and is very specific to each company. So understanding even if it is a requirement for your company becomes a pain. In this episode I share my experience going through the certification with Sastaticket and what I learnt during that.

In order to write faster code you need to understand (or profile) your current code and visually or numerically breakdown the pieces which need improvement. How do you do this? We use code profilers! In this episode we dive into the world of tweaking code performance and getting the most out of the hardware its running on.

Mono repos is not a new concept - but one often visited by growing organisations to help reduce the overhead of managing thoses dozens of repos in the company. Is it better? Well that depends on the company and how they work. In this episode I go through the differences and reasons why one would consider a mono repo over a multi repo.

Now a days even a simple CRUD based app can have a lot of moving parts - esp if deployed with something like Kubernetes. Our complex applications are growing even faster, interacting with many different libraries and handling greater and more variable loads. How do we get insight into whats really happening on the code level any more? Long gone are the days when a simple CPU usage alert would be enough to monitor most applications - today we need to monitor on different level and metrics to get a better picture of whats going on. This and more in this episode.

Many thanks to the GDE program for creating this initiative! In this episode I cover my learnings from buying my first "professional" mic - what are some of the things you need to know and how is my experience going so far with the mic I have bought.

Typescript is becoming one of the popular languages to use - but do we really need it? How helpful are the typings we write? What makes it stand out from Javascript? This and more in this episode covering the use of Typescript, benefits of typings, and some things to look out for.

As Docker and Containerisation have taken over the world - we have many different options to deploy our apps to. Within the AWS ecosystem, we have a few options from EKS, ECS and ElasticBeanstalk. While the latter might sound like an outdated option in front of EKS and ECS, it is surprisingly useful - this and more in this weeks podcast.

Gone are the days when we need tech teams to setup a simple website. Now a days we simply sign up like we do for Gmail and have everything setup in no time! And it costs just a few dollars per month. No Code tech is awesome and its here to make those mundane development roles redundant. With loads of new tools out there from making simple websites to building complex data entry apps or dashboards - the ecosystem is moving fast to empower the non-technical folks out there so build tech is more accessible and it becomes easier to innovate.

Writing documentation is not something we think about every day - but it is something we are interacting with on a daily basis. In a world of ever changing requirements and technology, if we are not documenting our technical journey properly - we are simply rebuilding the same wheel again and again. Documentation is at the core of any product, how its build, why it's built that way. And in this podcast we cover how we write this documentation in the best way possible. This is a summary of the technical writing course linked below.

Technical Writing Docs: 

https://developers.google.com/tech-writing/one

Register for workshop here:

https://docs.google.com/forms/d/e/1FAIpQLSc26RofABVtJkF1gRE7Pm7u2RBd6NBRbdqtYVUxYD_DVKwVUg/viewform

Writing async code is important when building any application. You dont want your threads to be blocked on http calls and db look ups - esp if the language you are using is single threaded! I was surprised to find that most code written in Python is synchronous though that is now rapidly changing. Writing asynchronous code right now means using libraries like Celery which do all the work in a separate process. What is it, how do we use it and how does it work - this and more in this episode.

In this session, I collect all the tips and tricks I have been using over the last decade and bring them together for a short 10 min checklist for making your website super fast. You can bookmark this episode and run through the points every time your website needs a boost!

If you are building a services company or an e-commerce store in Pakistan - eventually you will realize you need to accept online payments. Normally this requires a Payment Gateway - a simple service you can sign up to and integrate in a matter of minutes into your platform. But its not that easy in Pakistan as yet. What are the options? What works and what doesnt work? All this and more in this weeks podcast.

Micro-services is the popular architecture pattern which has taken the world over by storm. The promise of small manageable applications with fast iterations and seamless efficient scaling  has attracted many developers on taking on this as a default pattern for all their apps. But we should not be too quick dismiss the handy old monolith. In this episode we take a dive into micro-services to see if they are really that amazing underneath.

Continuous Integration (CI) is slowly becoming one of the main pillars for software development. As our software becomes more complex we add more and more tooling around it to automate, so we can focus on whats important - writing code. CI is what runs most of that tooling and keeps it out of the way.. until we need break something and it needs our attention. It gives us the confidence to deploy, again and again so that we move faster in this rapidly changing world.

Serverless architecture has become very popular over the last few years. In this episode we cover an introduction to this topic and talk about how Cloudflare Workers have implemented the serverless architecture. I share different ideas and thoughts on the ideology behind Cloudflare Workers and when they are most useful.

I recently rebuilt my personal website (imars.info) using NextJS and in this episode I share my experience in building it and my thoughts on the framework and how it compares to Wordpress which is something I have been using for a long time.  _This is not sponsored content_

AMP enables the creation of websites and ads that are consistently fast, beautiful and high-performing across devices and distribution platforms. It is an opensource toolkit which can be used for static pages as well as dynamic ones. But why do we need AMP? And how does it make our web faster? This and more in this short introduction of AMP HTML.

In this final episode of the Core Web Vital series, we bring together all that we have learnt and talk about the thing that matters the most - the 3 metrics set by Google which show how good or bad our website is currently performing. Largest Contentful Paint, First Input Delay and Cumulative Layout Shift - what are they, how are they calculated and how can we improve them? Just a few of the things I will share in this episode.

How does a browser render an HTML page? In this episode we dive into the world of DOM, CSSOM and Critical Render Paths. We understand what makes our website loading slow and how can we remove it so it can be made faster.

Core Web Vitals is the new standard set by Google to measure performance of our websites. In this multi-part series we will be covering on different aspects of web performance, starting with the basics. What makes a slow website and what are some of the things we need to keep in mind when reviewing our own website. In the upcoming episodes we will discuss what makes the website slow (technically) and how Core Vitals help us make it better.

In this episode I cover another very popular question I get - what kind of mobile app should I be building? With the rise of PWAs, Hybrids, Compile to Native, Flutter and native apps one does get lost in trying to figure out the best course of action. I try to explain the high level pros and cons for each and what really matter when making the decision for your mobile app.

In the previous episode we talked about what is the whole point of design systems, and if we need them or not. In this episode we cover a more technical perspective and discuss how do we go about building a design system from a flow perspective. We will also discuss some tools which can help us build this as well.

Here is another over hyped term. Every company seems to be making one now a days? But what is a design system really? How does it help you create better frontends? Is it really more than just a component library? Will be answering this and more in this episode.

Another popular question I get asked is which language is best to start off with when learning programming. Here is a quick overview of the two languages, their similarity and differences and a quick discussion on what are something of the things to think about when choosing one of them.

The most controversial topic in the frontend web space - which framework is better and which one should we spent time learning? The answer is not that straight forward! - I review both of them and share my perspective on what makes each the best.

One of the most frequently asked questions to me is on how does one become a GDE, and the next most popular is what is a GDE :) - so I have spent a few minutes explaining them both here!

Welcome to Karachi Wala Developer Podcast! In this series, I will share a little about my experience about different technologies and things related to startups. My name is Mashhood and currently I am the CTO at Sastaticket, and Google Developer Expert for Web and Angular. I have been working since I was a kid, and now with the 10+ year professional experience under my belt, I will be using this micro-podcast medium to share my thoughts on different technical things.