DEV Community: Muthoni Hailu

Did I quit my journey? (Oh, and my new app, MobileMuuzaji)

Muthoni Hailu — Wed, 17 Jun 2026 04:46:50 +0000

Where have I been?

The simple answer...no. No I didn't quit.

But then, where have I been?

Existing mostly. I'm still going strong with my Calculus and Data Science course. I've even started learning French. Then I realized I'm just grinding my wheels in the dirt with no real progress. I actually came to that realization yesterday. After months and months of studying and thinking that I was upskilling, I finally decided to make use of an employment preparation tool offered by my university. It lets you test yourself with interviews and assessment practices. So, I took the first assessment,

and got a beautiful,

red-colored,

D. Fucking D.

I was in the 14th percentile. I have never felt like more of an idiot.

But then, there's something else I've been learning about this past month. It's called, the growth vs fixed mindset. All my life I've had a fixed mindset coz I never really had to try hard to pass. I have never gotten anything below a B throughout my school years. I've always topped my class. Now I'm in the bottom 14th percentile on a global scale. And I finally feel not so smart.

But I can't just quit. Not because I have this crazy, practically delusional willpower to make something of myself. It's simply because, I can't quit. The world I was born into doesn't allow that. I can't depend on my parents forever, I'd feel pathetic doing that. I can't depend on government handouts, my country is too pathetic to have anything that nice. I can't live the rest of my life with no income source, yayyy capitalism.

So it's either I do something about my shitty brain or die. And while not existing sounds very enticing sometimes, I'm not ready for that yet. I'm still a little curious about what the future holds. Not curious enough to live past 23 if my life doesn't turn around within the next few months. And I can't sit down and wait for my life to turn around, I have to actually do something about it. So, here I am, being mediocre but hoping that mediocre is at least good enough.

MobileMuuzaji

The main reason I decided to post today actually is to share a project I have built (mostly vibe-coded because I don't know the first thing about Kotlin) for my mum's business. There's a high likelihood I'll end up working there so I'd better start taking an interest in what happens, and figure out how I can make my Computer Science degree useful on a farm and small shop in the middle of nowhere in Kenya. Also, I love my mum.

I want to write a post about it because I was told this is a good way to get discovered by potential employers. Why did I start by writing a sob story that will most likely put off any potential employers? I don't know. I'm not that smart. I think we've established that already, haven't we?

This post aims to outline the development process and main features of MobileMuuzaji. This is a mobile POS (Point of Sale, not the other one you weirdo) terminal specifically designed for small business owners in Kenya. It allows them to easily track and manage inventory as well as employee details. It also offers data visualization tools so that they can model the benefits and drawbacks of their different inventory items (this is mostly here just to show off my understanding of data visualization :) ).

What Does MobileMuuzaji Actually Do?

Before diving into the analysis, here's a quick rundown of what the app offers at its current stage of development.

The core idea is simple: give market vendors and small shop owners a tool to track their stock and record sales without needing a constant internet connection.

Current Features

Authentication & Onboarding

A guided onboarding experience for first-time users with contextual tooltips that walk users through the app's key actions
Email and password signup and login with server-side validation
Persistent login sessions so users don't have to log in every time they open the app

Organization Management

Users can create and manage organizations (think of these as individual businesses or shops)
Organization admins can add employees by email and remove them from the organization without deleting their accounts
Both admin and employee roles are supported, with role-based access — for example, only admins can manage employees

Inventory Management

Add new products with a name, quantity, unit (supports both weight and volume units — kg, g, lb, oz, litres, ml, etc.) and cost per unit
Edit existing inventory items
Search inventory by item name in real time
Offline support — inventory changes made without internet are saved locally and automatically synced when connectivity is restored

Sales Recording

Record sales directly from the inventory list using a "Sell" button on each item
Two input modes: enter how many units were sold, or enter the leftover quantity and let the app calculate the amount sold
Built-in stock validation — the app prevents recording a sale that would exceed available stock
Offline support — sales recorded offline are queued and synced to the backend automatically

Sales Analysis

View sales history per organization
Filter sales by date (today, this week, this month, all time, or a custom date range)
Sort by date, alphabetical order, or earnings
Toggle between individual sales view and a grouped view that aggregates sales by product and shows total earnings per item

Offline-First Architecture

The app uses a local SQLite database (via Android Room) as a cache
All data is pre-loaded on login so the app remains fully functional without internet
Background sync via WorkManager pushes local changes to the PostgreSQL backend whenever connectivity is available
Offline-created items use temporary IDs that are replaced with server-assigned IDs on sync

SWOT Analysis

Let's be honest about where MobileMuuzaji stands right now — what it has going for it, where it's lacking, and what the landscape looks like.

Strengths

Built for the offline reality of its target market. A significant portion of Kenya's small business activity happens in areas with inconsistent mobile data coverage — markets, roadside stalls, semi-rural areas. Most inventory apps assume a reliable connection. MobileMuuzaji doesn't. The offline-first architecture means a market vendor in Gikomba or a small shop in Kisumu can still record every sale and track every product even when their data runs out or the signal drops.

Low barrier to entry. The app targets Android, which dominates the Kenyan smartphone market by a wide margin. There's no web portal to navigate, no desktop software to install — it's a single app that lives on the device the user already has.

Role-based access makes it suitable for small teams. A shop owner who employs a few staff members can give them access to record sales and manage inventory without giving them full administrative control. This is a meaningful feature for the target user who often delegates day-to-day operations to employees.

Flexible unit system. Supporting both weight and volume units might seem like a small detail, but it matters enormously in a market context where vendors sell everything from rice by the kilogram to cooking oil by the litre. A generic unit field wouldn't capture this accurately.

ACID-compliant transactions on the backend. Sales are processed as atomic transactions with row-level locking, meaning stock levels stay accurate even under concurrent load. This is the kind of backend reliability that many small-business tools at this price point (free) don't bother with.

Weaknesses

No financial reporting beyond basic earnings. Right now the app can tell you total earnings per product, but it can't generate a profit/loss statement, show trends over time with a chart, or export data to a spreadsheet. For a business owner trying to make informed decisions, this is a meaningful gap.

No price per unit tracking for sales. Sales currently record earnings as quantity × cost per unit, but there's no way to record a selling price that differs from the cost price. A vendor who buys onions at 50 KES/kg and sells at 80 KES/kg can't currently capture their actual profit margin — only their revenue.

No payment method tracking. M-Pesa, cash, and credit are all treated the same. In Kenya especially, distinguishing between cash and M-Pesa transactions is genuinely useful for reconciliation.

Employee management is online-only. Adding or removing employees requires an internet connection. For admin users in low-connectivity areas, this could be a frustrating limitation.

No data export. There's currently no way to export sales or inventory data to CSV, Excel, or PDF. For users who do their own bookkeeping or share reports with accountants, this matters.

Opportunities

Kenya's informal economy is enormous and underserved by software. The informal sector accounts for a significant share of Kenya's GDP and employment. The vast majority of these businesses have no digital tool for inventory or sales tracking — they rely on notebooks, memory, or nothing at all. The addressable market is large and the competition at the entry level is thin.

M-Pesa integration potential. Kenya has one of the world's most advanced mobile money ecosystems. An integration with M-Pesa's API would allow MobileMuuzaji to automatically reconcile sales against payments, which would be a genuinely compelling feature for Kenyan vendors and a strong differentiator against generic inventory apps.

Expansion to East Africa. The connectivity challenges and informal economy characteristics that make MobileMuuzaji relevant in Kenya are shared across Tanzania, Uganda, Rwanda, and Ethiopia. The offline-first architecture that was built out of necessity for the Kenyan market becomes a strategic asset if the app expands regionally.

WhatsApp Business integration. Many Kenyan small business owners use WhatsApp as their primary customer communication channel. A feature that lets users share stock availability or sales receipts directly via WhatsApp could meaningfully increase daily active usage.

Low-cost data analytics as a premium tier. As the user base grows, aggregated (anonymised) sales data could inform a premium analytics dashboard — showing seasonal trends, popular products by region, and demand forecasting. This could form the basis of a freemium revenue model.

Threats

Generic inventory apps from global players. Apps like Zoho Inventory, Wave, and various POS systems are available on the Play Store and some have free tiers. While they're not built for the Kenyan informal market specifically, they have significantly larger feature sets and established brand recognition. MobileMuuzaji needs to compete on local relevance and simplicity, not features.

Smartphone upgrade cycles. The app targets Android, but older, lower-spec Android devices are common in the target market. Performance on low-end hardware needs ongoing attention to avoid a poor experience for users on devices with limited RAM.

Data privacy and trust. Kenyan business owners may be cautious about storing sensitive sales and inventory data on a third-party server, particularly smaller vendors who operate informally. Building trust — through transparency about data handling, local hosting options, and clear privacy terms — will be important for adoption.

Connectivity improvements cutting both ways. As Kenya's mobile data infrastructure improves (5G rollout, cheaper data bundles), the offline-first differentiator becomes less compelling. The app needs to continue adding value in ways that remain relevant regardless of connectivity.

Where do we go from here?

If you remember from my intro I mentioned that I wanted to have data visualization tools in my app. Currently, I don't, and as a data scientist to be, this fact really bothers me. So in the next version of MobileMuuzaji, that will be the first addition (not just data visualization but data management tools as well). The main data management tools I am targeting are:

A data visualization feature allowing customers to see trends in how their different inventory is performing
Data download features with support for csv, json and xlsx file formats making it easier to export data and work on it elsewhere
Potentially (this is a big thing, I don't even know if I'll manage) a mobile-based Jupyter Notebook integration so data can be analyzed and manipulated directly from inside the app

An additional feature alongside the data management will be adding the ability to manage purchases as well and map out profits. This tool will be especially useful, since, as every business owner knows, what matters is not the gross but the profit. Possible integration with VAT management would be a bonus.

But as of now, my mum loves the application, so that's what she will get. It is still being used at a small scale and has not been fully deployed for widespread use, but when it does, I'll update whoever was bored enough to follow my story.

Check out the github here

Another week, another small step forward

Muthoni Hailu — Sun, 12 Apr 2026 11:43:54 +0000

There's honestly not much I can report on this week. In fact, I am thinking of switching to making updates on here once every fortnight instead of weekly.

This week, I was able to finish:

Khan Academy's Statistics and Probability: Displaying and comparing quantitative data - Describing and comparing distributions
Complete my client's backend

I did not make a submission to the competition I had planned for (Kaggle's TPUs + Cassava Leaf Disease competition). I realized I should have started with Petals to the Metal which is an introductory competition but I gave up on doing that after like my hundredth "Your notebook used all its memory" error. So that will be my main focus this week, not the completing the competition bit but not giving up because of error messages.

Working with a freelance client is still a pain in the ass but its not like I have an option. I also really struggled fighting my depressive tendencies but I was still able to finish all this which makes me feel kinda happy.

The plan for the next two weeks involves:

Khan Academy's Statistics and Probability: Analyzing categorical data - Review
Khan Academy's Calculus: Differentiation: composite, implicit, and inverse functions - Review Differentiating inverse trigonometric functions and Selecting procedures for calculating derivatives: strategy
Make a submission to Kaggle's Petals to the Metal competition
Try to get my client to respond on what type of chatbot they want.

That's all for this week. Bye.

The most important lesson I've learnt as a dev so far

Muthoni Hailu — Sat, 04 Apr 2026 14:58:48 +0000

I am mediocre.

I am not the smartest, nor the fastest, nor the most passionate. Being the best student in my high school class meant nothing. There are millions of other top students all across the globe, probably thousands just in my country. In comparison to them, I'm just average. And maybe that's ok.

Maybe I don't have to have landed a FAANG role a few weeks post my final university exam. Maybe its OK that there's a likelihood that I will graduate with a second class honors and not a first. Maybe my fear of mediocrity is the reason why I have not done anything at all, let alone something mediocre.

This week I was struck with this blatant fact while revisiting my Khan Academy Calculus work. I had to actually put in effort. I failed multiple quizzes and had to go through the material multiple times to pass. I've never had to do that before. It's a painful experience realizing that the one thing you had pride in, your intellect, is actually nothing special. But I'm glad I worked through it. I'm glad I didn't give up. I'm glad I cried yet got up the next day and kept practicing. I'll probably still be jobless 6 months from now. Considering the job market, I'd be extremely fortunate to land a role (not freelancing) before the end of the year. But if I don't give up, I'll still be content with that, hopefully.

As mentioned, this week I revisited some differential calculus work. I completed the Statistics and Probability work I had set out to do in my previous post as well as Kaggle's Computer Vision course. I was also able to build the frontend for my freelance project (I am realizing the toughest part of freelancing is working with the client and not really the development. Though, the fact that I am so grateful for the job kinda helps me through that)

As for this coming week, my goals are:

Khan Academy's Statistics and Probability: Displaying and comparing quantitative data - Describing and comparing distributions
Make a submission to Kaggle's Getting Started: TPUs + Cassava Leaf Disease competition
Complete my client's backend

That's all.

First Week's Progress

Muthoni Hailu — Fri, 27 Mar 2026 20:01:21 +0000

So, the first week of my attempt is coming to a close and things are much better than I expected them to be actually.

But first, I'll start with the boring bit. This week, my main focus was supposed to be progressing through Khan Academy's Statistics and Probability course by completing one unit. Surprisingly enough, I did just that. I was even able to do a bit of review for the Calculus course that I "finished" last year (I finished in the sense that I got to mastery level for all units, mostly by using Google Image search to answer the questions for me. Why would someone do that when they actually want to learn? I guess I just value the dopamine spike you get from that little tune that plays when you finish a level over actual learning. So now I have to go all the way back like someone who's never stepped foot in a Calculus class). Tangent over, back to Stats and Probability.

The course was rather simple, too simple actually. I am scared because I know that eventually things will get tougher. What if I'm unable to keep my motivation then? The only place I struggled was with identifying trends in categorical data. I guess my logic skills are absolutely awful. But for some reason, for the first time ever, I am motivated to practice and get better. I have chosen to use a random worksheet that I found online. I'll make an update on my progress next week.

As for this coming week, I plan to work on the next unit which is displaying and comparing quantitative data. But, I will avoid a pitfall that I have made previously which was assuming because the first unit was too easy, that I can rush through all the other units at a similar place. So, I will only aim to complete at least the first subsection (it seems to have a lot of quizzes so I may take some time on it).

Next, I had planned to work on DSA, focusing on heaps. Then, at the start of the week, I randomly decided to focus on Computer Vision instead and work on a new project idea. I did finish four units in Kaggle's CV course and made the frontend for my project, which is really cool. Then I went to see a dentist, and this is where things changed.

My dentist happened to ask what I was studying. When I told her I had just completed a CS course, she randomly offered me a gig building a website for her. I am still in shock. If you had told me, as I was sitting my final exam, that I would end up with a freelance gig just a week later, I would have slapped you probably (I may be a bit too violent now that I think of it). But, here I am. Granted, this gig won't make me much money and my experience is definitely an inaccurate depiction of the current freelance market, but its something. And it gives me hope.

However, I am panicking. My hope is that I can make a website so good that she'll recommend me to more of her friends who are seeking a similar service. But, what if my site is shit? I've never built anything professional before. Heck, I've barely built any substantial project outside of the school ones that I vibe-coded and halfassed. I am scared. But its not like I can back out now. This change does put a pause to my study plans for a bit though. Yes, I'll continue studying, but I'll be going at a much slower rate.

Based on all that, here's my plan for next week:

Khan Academy's Statistics and Probability: Displaying and comparing quantitative data - Displaying quantitative data with graphs
Kaggle courses: Computer Vision - Custom Convnets and Data Augmentation
Work on the frontend for my dentist's website

That's all for now. If I was a praying man I'd be on my knees right now because I am scared as hell. But oh well...

Attempt at a new beginning

Muthoni Hailu — Fri, 20 Mar 2026 14:09:04 +0000

Hello, my name is Muthoni.

I am 22 years old and I just did my final university exam 4 days ago. I was taking a BSc in Computer Science from the University of London distance learning program but despite that, I feel like I have no programming skills. I have no portfolio, I barely remember high school math, let alone having a college level understanding and I am basically unemployable.

This is through no fault of the university (except for the math part, they do not offer much when it comes to that). It is my fault. I have spent the past three and a half years mostly bedrotting and being depressed while my peers participated in hackathons, contributed to open source projects and built portfolios through freelance work. I have tried to get my life together multiple times during those three and a half years. I have started and quit multiple fitness journeys. I have vibe-coded a bunch of shitty projects that I am not proud of, and have started and quit multiple attempts at completing Khan Academy's math courses.

Well, what makes this attempt different? Nothing. Maybe the fear of becoming homeless because now I have finished university and my parents will be expecting me to get a job and move out soon. But even that prospect doesn't seem to scare me enough to get my life together. I don't know if this attempt will work. I don't know if this will end up being a dead blog after three weeks and I'll go back to my miserable reality. But I can't afford to not to do anything.

So this blog will mostly detail my progress. Currently, I am mostly focused on learning AI. I will be following the AI roadmap guide by Jean, a Stanford graduate. I am starting from the very beginning with the Math courses. I hope I'll be done by March 23rd 2028. I will be making updates every week, detailing the progress I have made, any insights I got and my plans for the next week.

The first week for me starts on Monday, that's March 23rd. My main focus will be learning Probability and Statistics and making some Data Structures and Arrays progress. I will try and finish:

Khan Academy's Probability and Statistics Course: Analyzing one Categorical Variable - The first unit
Learning Heaps Data Structure by pooling from various online resources

That goal is small, but hopefully it is small enough that I won't feel the urge to quit after two days.

I don't expect anyone to read this, I mostly just want something to look back on and see how far I've come if this is successful, or feel shame about if it isn't. If you've actually read this, I'm glad you took interest but I'm also really scared about you continuing to follow my progress in case I fail. But I can't stop you from following along, can I?