DEV Community: Leapfrog Technology

Bridging Discovery to Design: A technical blueprint checklist

Leapfrog Technology — Mon, 28 Jul 2025 09:26:04 +0000

Written by Praneen Shakya, Solution Architect at Leapfrog Technology

The journey from discovery to design in any product development process is a critical phase that ensures success. One of the most overlooked yet essential steps in this process is creating a Technical Blueprint Checklist. This checklist acts as a bridge between understanding user needs and implementing feasible, well-designed solutions.

This blog provides an overview of what a technical blueprint checklist is, why it is required, key elements in the checklist, and a sample checklist to guide you through the Discovery and Design journey.

Technical blueprint checklist for scalable software

A technical blueprint is the software equivalent of an engineering diagram for buildings. It helps you navigate the solution’s architecture and its intricate details during execution.

A technical blueprint checklist is a carefully curated list of items that we shouldn’t overlook while composing a technical blueprint.

Let's take an example of building a house. An engineer designs the rooms, adds electrical wiring, walls, doors and windows, but forgets to account for the wall switches.

You might think that this will not create much trouble as electricians usually remember to install switches when laying out wiring. However, even a single oversight like this during execution can create unexpected chaos and complications.

Now, imagine a similar mistake happening at scale: ten tweaks across ten different areas will create chaos and inefficiencies. This is why most of the big buildings get abandoned due to lack of meticulous design.

The stakes are even higher in case of software development because software products are virtual rather than physical. Every detail in the technical blueprint must be precisely accounted for to avoid software issues.

That’s why we need to have a comprehensive technical blueprint checklist to ensure every detail is captured during the Discovery and Design journey. The checklist will prevent things from being missed out ensuring scalable and reliable software development.

Must-haves in your technical blueprint checklist

Functional and nonfunctional requirements

Clearly define what the system should do (functional requirements) by specifying its features, behaviors, and capabilities to meet user needs. For non-functional requirements, define how a system should perform, focusing on quality attributes like reliability, scalability, performance, and usability.

Data flow diagrams

Include a data flow diagram which maps out how input data flows from one step to the other, gets processed in the way, and ultimately ends up with the output.

ER diagram

Visualize the system’s data structure—what entities exist, their attributes, and how they relate. This is crucial for designing an efficient, scalable database.

Story breakdown and prioritization

Break user stories into smaller, manageable tasks and rank them according to the importance and urgency to maximize value delivery and to align with project goals.

User journey

Visualize step-by-step interactions and experiences that a user goes through to achieve a specific goal within a system, over a period of time.

C4 model

Use this framework to visualize software architecture across four levels: Context, Container, Component, and Code. Detailed elaboration on C4 can be read in my other article here.

High-level architecture diagram

Incorporate an overview of a system’s structure, showcasing its main software components, their interactions, and external integrations without delving into implementation details.

Cloud architecture diagram

If your software uses cloud, illustrate the components, services, and interactions within a cloud-based system, including servers, databases, storage, networking, and integrations with cloud providers.

Architecture diagram

Document key architectural decisions - context, rationale and outcomes- made during a project’s development. Icepanel is a great tool that combines the C4 model, workflows, and acts as a long-term ADR.

API design

Define the structure, endpoints, data format, and behaviors of an API. Ensure it is intuitive, efficient, and aligned with the needs of the users and system architecture.

Proof of concept plan

Having a proof of concept plan will help you validate the feasibility of a solution or idea. Focus on validating key technical aspects, assumptions, or concepts with minimal resources and effort.

Minimum viable product plan

Have a minimum viable product plan which is the simplest version of a product that includes only the essential features necessary to satisfy early adopters and validate assumptions, allowing for feedback and iteration.

Cloud component capacity planning

Estimate and manage the required resources (such as compute, storage, and network) to handle expected workloads, traffic, and growth while maintaining performance and cost-efficiency.

Infrastructure cost estimation

Predict the financial costs associated with the resources needed to run an IT infrastructure, including compute, storage, networking, and services, often factoring in scalability, usage patterns, and cloud provider pricing models.

Task estimations

Estimate the time, effort, and resources required to complete a specific task or set of tasks, typically used in project management to plan and allocate resources effectively.

Staff plan

Outline required personnel for a project or organization, detailing roles, responsibilities, skills, and the timeline for hiring or allocating staff to ensure the project's successful execution.

Project timeline

Create a visual representation of the project’s schedule, outlining key milestones, tasks, and deadlines, helping to track progress and ensure timely delivery of deliverables.

Get started with a sample checklist

With over 10 years in the software industry, I’ve helped uncover requirements for hundreds of clients, each project revealing unique needs but often centered around a few core requirements. Identifying these, exploring them, and crafting a roadmap to execution defines the Discovery-to-Design journey.

The checklist below provides a sample to start your Discovery and Design journey, which can be customized as per the project scope and requirements.

To explore more insightful blogs, visit www.lftechnology.com/blogs

How to be a successful Product Manager for AI products

Leapfrog Technology — Tue, 27 Aug 2019 08:12:40 +0000

You can’t exist in the tech world today without hearing about AI at every turn. The business value is clear— automate tasks that would take humans far longer and gain insight into valuable new data. Given that our product managers regularly use data to support their hypotheses about new features, where do they fit into this AI equation? How can they overcome any knowledge gaps and lead successful product launches utilizing machine learning?

The answer is not as complicated as it may seem. A product manager with a solid grasp for the tactical process can rise to the challenge.

What product managers need to know about AI

We know that product managers need to think tactical, but how deep should their AI knowledge go? Ultimately, product managers just need to know what AI can do, and conversely, what the limitations are. This is important in determining whether AI is even a viable solution for the feature in question. Knowing what data is available, and how it can further AI development allows the product manager to tactically steer the product.

Taking it a step further, the product manager should be able to know how AI will integrate into the current product, and how to improve the AI models moving forward. Do the collected data aid future initiatives? What other data sets could be useful for future AI capabilities in the product? These are questions that can help propel efficiency in the machine learning model. The key is to not get bogged down with implementation level questions and stick to the high-level strategy.

A good exercise is to test your knowledge of useful applications of AI by thinking about how it can aid your own company. A good place to look is at your rule-based systems that depend on certain classifications. For example, here at Leapfrog, we came up with our management of employee supervised learning. This system depends on classifications such as whether people are high or low performers, their total length of employment, and employee feedback. We can use this to determine goalposts of what employees can achieve in 3 months, 6 months, or in one year. Continued data and feedback over time will grow these models. Ultimately, we can strengthen company culture, morale, and growth by investing AI into our employees. Looking at simple internal solutions allows us to start flexing our machine learning muscles.

How to structure AI products

A product manager can use their strategic knowledge to structure an AI project appropriately. The following high-level steps can guide the flow necessary for an AI project.

Get the data: Some clients may desire an AI solution, but will not have the tools in place for this to be viable. Without the right set of customer data, you will not be able to strategically apply it to an AI problem.
Explore and analyze the data: How can I use the data? Does this data support my product hypothesis? What initial findings do I see? The product manager can engage with the data at a high level to see if this fulfills the needs of the feature in question.
Validate the hypothesis with users: As with any project, it is imperative that we validate our assumptions with our user. Testing our initial ideas will allow us to make valuable iterations throughout the product cycle. By starting small and iterating, product development can gain speed.
Create quick and dirty models: The team can use the initial research and hypothesis to begin building quick models. More questions will arise, but don’t lose sight of the initial hypothesis and goal.
Tune the models: Iterating on the initial models will continue to confirm or deny the original hypothesis. We can ask ourselves, “What is the minimum level of functionality that is acceptable?” Keep tuning this model until solid results are delivered.

Ask the right questions

At the end of the day, the biggest measure of success will be if a product manager can ask the right questions at every stage of the product cycle. They are steering the features from hypothesis through development. Success in machine learning projects is ultimately measured similarly to regular development projects. If the product manager can retain their tactical approach, keep the team aligned around key product goals, and stay focused on the experience of the end-user, then the project will be in good hands.

Want to take the next leap and learn more about integrating AI in your product?

Evolution of AI

Leapfrog Technology — Tue, 20 Aug 2019 11:06:04 +0000

In the past few years, AI evolved into a powerful tool that enables machines to think and act like humans. Moreover, it has garnered focus from tech companies around the world and is considered as the next significant technological shift after the evolution in mobile and cloud platforms. Some even call it the fourth industrial revolution. Forbes states, “By 2020, businesses that use AI and related technologies like machine learning and deep learning to uncover new business insights will take $1.2 trillion each year from competitors that don’t employ these technologies.”

This article gives you an overview of the evolution of AI and sets a foundational understanding of important milestones that led the path for AI surge.

Artificial Intelligence (AI)

According to the Merriam Webster dictionary, “Artificial intelligence is a branch of computer science dealing with the simulation of intelligent behavior in computers.” When a machine can make intelligent decisions, it can be referred to as being intelligent- artificially. We mostly see people using the terms of machine learning, deep learning, and AI synonymously. However, deep learning is a subset of machine learning, and machine learning is a subset of AI.

A Venn diagram showing how deep learning is a kind of representation learning, which is, in turn, a kind of machine learning, which is used for many but not all approaches to AI. Each section of the Venn diagram includes an example of AI.

When did the AI surge begin?

Back in the 1800s, AI was limited in myths, fiction, and speculation. Classical philosophers envisioned machines integrated into human beings. However, they were just portrayed in fiction work like Mary Shelly’s “Frankenstein” then. The real initiation in AI began in 1956. The seed that led towards an AI future was a workshop in Darthmod College, attendees of which were claimed as AI leaders for decades to come.

The AI surge began with six major design goals as follows:

Teach machines to reason in accordance to perform sophisticated mental tasks like playing chess, proving mathematical theorems, and others.
Knowledge representation for machines to interact with the real world as humans do — machines needed to be able to identify objects, people, and languages. Programming language Lisp was developed for this very purpose.
Teach machines to plan and navigate around the world we live in. With this, machines could autonomously move around by navigating themselves.
Enable machines to process natural language so that they can understand language, conversations and the context of speech.
Train machines to perceive the way humans do- touch, feel, sight, hearing, and taste.
General Intelligence that included emotional intelligence, intuition, and creativity.

All these goals set the foundation to build a machine with human capabilities. Millions of dollars were invested in bringing their vision to life. However, soon, the US government realized the absence of powerful computing technologies needed to implement AI. The funds were withdrawn, and the journey took the first halt in the late 80s.

The need for a massive amount of data and enormous computing power disrupted the progress in the 80s. The 21st century, however, brought the concept quickly back to life proving Moore's law. The heavy processing power that tiny silicons hold today has made AI feasible in the current context, also enabling to build improved algorithms.

There have been four successive catalysts in the AI rebirth and revolution:

The democratization of AI knowledge that began when world-class research contents were made available to the masses- starting with MOOCs from Stanford University with Andrew NG and Intro to ML by Sebastian Thurn and Katie Malone from Udacity.
Data and Computing Power (cloud and GPU) that made AI accessible to the masses without enormous upfront investment or being a mega-corporation.
Even with access to data and computing power, you had to be an AI specialist to leverage it. However, in 2015, there was a proliferation of new tools and frameworks that made exploring and operationalizing production-level AI feasible to the masses. You can now build on the backs of giants like Google (Tensorflow), and Facebook( PyTorch). Numerous organizations have been founded with the democratization of AI like FastAI and OpenAI.
In the past two years, AI as a service has taken this a step further, enabling easier prototyping, exploration, and even building sophisticated and intelligent use-case specific AI's in the product. There are platforms like Azure AI, AWS AI, Google Cloud AI, IBM Cloud AI, and many more that provides AI as a Service.

Humankind has witnessed AI identifying cat videos to the invention of self-driving cars. More companies are trying to apply AI to solve problems and build an AI strategy in the organization. The journey of AI, which began with six goals, is gradually being achieved today. With further advancement in technology, a future with humans and machines living together seems possible. Now, we can either choose to observe the trend or start planning our AI strategy to create an impact in a world driven by technology.

Thank you for reading!

If you want to learn more about implementing AI into your Product

Introducing Chill: An OpenSource Service Monitoring Tool

Leapfrog Technology — Mon, 17 Jun 2019 04:08:49 +0000

It's been a while since you've heard from us about Chill. Our last update was nearly two years ago. Phew! All of you in some point of time might have felt the urge to create something new only for the excitement to slowly fade away. We went through a similar situation ourselves. 😉

Here's the backstory, a few months back, we figured out that the tool we built was up and running in production for more than a year. It was still monitoring some of our production websites without any problems. We were not amazed as we would get downtime notifications from Chill on our Slack channels from time to time. This simple "ping tool" as some like to say, was performing very well without much supervision. That is why we decided to take it further by adding in new features and sticking with the minimal approach that we took when we started the project.

It's 2019, and here are some of the changes that we have made to "Chill" - our minimal and dependable service monitoring tool.

Stable Release
We quietly released our stable version of Chill v1.0.0 on Mar 12, 2019. Since the release, we have been making quite a few changes to Chill and its internals.

A New Logo
The very first change to Chill is our new logo. We always wanted Chill to have this sense of relaxation as it kept an eye on our services. The word "Chill" in Nepali translates to an Eagle, which is a perfect mascot for keeping good sight of our services. No wonder we have an eagle wearing cool shades as our logo. The mixture of both the Nepali and English meaning to Chill brings the logo to life. This is precisely how we got the name "Chill".

New Status Dashboard
We have revamped our status dashboard and included a few more items in there. First of all, we have included a new section called "Past Incidents" just below the status report. This section of the page includes any past events or downtime information of different services grouped by date. Similarly, the design of the whole page has been tweaked a bit to make it look minimal and clean.

Chill Status Dashboard

Support for HTTP Basic Authentication

Chill 1.1.0 adds support for HTTP Basic Authentication. In previous versions of Chill, your website/application would be marked as down if it was behind an HTTP Basic Authentication. Now, you just have to update your Chill configuration and provide your username and password for Chill to start monitoring your website behind an HTTP Basic Auth.

Support for Under Maintenance Status
Chill is now aware of websites under maintenance. If your website is under maintenance and is reporting an HTTP status of 503 with retryAfter header set to a value more than zero, Chill reports its status as Under Maintenance. This is particularly useful for web applications and services when they are on scheduled maintenance and don't want Chill to report downtime.
The integration of Under Maintenance status on the dashboard page is yet to come.

Service Retry Support
To make sure that the website is down, we now attempt to fetch the status of the website multiple times before throwing in a notification. This entirely removes any false alarms that we saw in previous versions of Chill. The time interval for service retry can be changed in Chill configuration as well. We have sane defaults in place for those who don't want to change the intervals.

Internal Changes

All our packages and dependencies are in the most recent version.
We have merged all our repositories to one single repository. This allows us to collaborate and make changes to Chill easily. Yes, installing Chill is still a pain, but we are soon going to make it easier than ever.
Fixes to broken Websocket integration.
Removal of Hipchat notifier. So long Hipchat! 👋

The Future
We are not done. We want to add more features to Chill while staying true to its core. We will continue to make steady changes to Chill in the coming months. Some exciting new additions that are coming to Chill are SSL Status Reporting and Slackbot Integration. In regards, we want to make sure that we write an update blog for every release. 😎

Stay up-to-date with our releases at https://github.com/leapfrogtechnology/chill/releases

Want to contribute to Chill? Want to send us feedback? Feel free to reach out at chill@lftechnology.com

Setting up your Mac for Machine Learning

Leapfrog Technology — Mon, 17 Jun 2019 03:58:37 +0000

From zero to running the first Machine Learning and Deep Learning projects within hours using scikit-learn, Keras, and TensorFlow

I bought a new Mac. I need to install everything necessary to build the Machine Learning project on my local machine. I could have copied everything from my old Mac using Time Machine. However, I wanted to start from the scratch so that I can document the steps and I can share with all of those who daily job is not coding but who wants to get hands dirty now and then.

I am not a developer. I am a Product Manager. I don’t code daily, but it’s fun to write some code when necessary. At least, I can clone a project locally in my machine.

Install Python
Check whether there is Python installed by default. Mac comes with default Python version 2.7.0. I will use Anaconda for Python and different packages for Data Science and Machine Learning.
Install Miniconda
I chose Miniconda because I don’t need all the packages right away. I can install those packages in the virtual environments I create on the need basis. I like to isolate dependent packages for different types.
- Anaconda vs Miniconda vs Virtual env
- How to install Miniconda
Managing environments
- Why you need environments
- Conda documentation
Install Jupyter Notebook
Install Jupyter Notebook
How do you install packages in Jupyter notebook? There are complexities we need to be careful. If you are interested, then here is a good article.
After you install Jupyter Notebook, you can follow this notebook to install and check all those packages.

Hope you find this curation useful!

Now that your Macbook is ready, want to take the next leap and learn more about integrating AI in your product?

This article was first published on lftechnology blog