Olawale Badekale

Machine Learning Engineer | MSc Computer Science | Building Production AI Systems | NLP | Computer Vision | MLOps

[Olawale Badekale @cuthbert_ola]

Tuchel returning to Germany after sabotaging England’s World Cup

[Johnmark Obiefuna @jayhemz]

In case you've been seeing my tweets, note this carefully.

I'm not the best in any single thing, but I'm the best in the composition of things.

This perspective is useful in case of those struggling about not knowing where to focus on. Early in my career, I start a tutorial in Data Analytics, stop, enter DevOps, stop, enter programming, stop, back to data analytics.

I sat myself down one day and wondered how I'll "make it" laidiss.

I HAVE NOT YET MADE IT, but the direction is clearer. I operate at a high level, a strategic level, bird's eye view if you will. And in some cases, I've been flown in by clients locally and internationally to fix raging problems, which I've mostly been successful doing.

So in case you're confused, it's normal. Get all you can get in different, almost unrelated areas, you may end up being the glue when unrelated challenges show up.

Put yourself out there, and keep pushing.

[Adéwálé @__adewale]

Dream team 💙

[Bola Banjo @bolaabanjo]

AI engineering is really hard.
I didn’t even study this much in all my 5 years in school.

[Chisom Nwokwu🌟 @tech_queen]

Towards the end of 2025, I made a decision that felt uncomfortable but necessary: step away from social media for 3 months, LOCK IN, and focus deeply on the life I’m trying to build.✨

Three months later, I can honestly say it was one of the best decisions I made.

The past few months were pretty quiet. Boring, even. But behind the scenes, I:
- Started and completed projects I'm genuinely proud of,
- Upskilled aggressively,
- Landed some MIND-BLOWING opportunities,
- Spent more alone time with God and
- Got really intentional about the rest of my year.

Sometimes the most productive seasons of your life looks like silence from the outside❤️

I'm back now, and I have a lot of good news to share🎉

But first... what did I miss? 👀

[Oba Akute🐎 @Vynchie_]

House of Assembly 🎹

[Johnny Minerals @mineralsfc]

#BlueCoOut

[Asanwa.sol @Chizitere_xyz]

Modern women obsession with "soft life" and "princess treatment" has made a whole generation of women completely blind to the brutal reality of what it actually takes for a man to provide that.

You are demanding a man be endlessly romantic, available 24/7, and constantly planning aesthetic dates, while he is literally in the trenches fighting a ruthless economy to build the very empire you want to rest in.
​You cannot demand the spoils of war and then complain that the soldier is too tired to entertain you when he finally gets home. We want the absolute financial security of a conqueror, but we expect him to have the carefree personality and infinite free time of a golden retriever. A man cannot simultaneously be at peace and at war. If he is building your fortress, you have to give him the grace to be exhausted

[AKIN-MADE @bibzyCarter]

I work hard everyday so my daughters don't have to tweet like this.

[Kharis @kharis_micheal]

Your girlfriend has a guy that she specifically texts or calls after you upset her.

[Heart Mattaz by Rita @HeartMattaz]

Marriage needs structure both mentally and financially. This is why you must marry someone who doesn't see her money as her money and your money as the family money.

Someone willing to be in a marriage with you 1000% in all ramifications. Someone who understands when to cut costs and stay within family cash flow.

Someone with depth, empathy, and selflessness. Someone you can lean on and someone who wouldn't use your financial weakness against you.

Two good heads are better than one.

[Olawale Badekale @cuthbert_ola]

One of the most important rule of ML is knowing the https://t.co/obNL6uj4TV matrix multiplication rule.

(A, B) dot ( B, C ) = ( A, C)

Where A is the sample(input)
B is the feature(input)
C is the output

[Deep-ML @real_deep_ml]

Not sure who needs to hear this but stop overcomplicating the learning process,

If you want to learn ML stop going on reddit or X or whatever looking up “how do I learn ML” to quote shai labeouf JUST DO IT, find an interesting problem (not mnist unless you really find classifying numbers super interesting) and build it get stuck do some research on why you are stuck and keep building (if you are using chat ask it not to give you code, chat is helpful but if it just writes the code for you you won’t learn anything, read the reasoning and try and type it your self)

If you are spending hours coming up with the perfect learning path you are just kidding yourself, it is a lot easier to make a plan then to actually study/ learn (I did this for a while, I made a learning path and a few days in I was like no I need to add something else and spent hours and days making a learning path to run away from actually doing something hard)

Ultimate guid to learn ML:

1.⁠Find an interesting problem (to you)
2.⁠Try and build it
3.⁠Get stuck
4.⁠Research why you are stuck
5.⁠back to Step 2

[ₕₐₘₚₜₒₙ @hamptonism]

I’m addicted to learning because the right information can change your life.

[Harper Carroll @HarperSCarroll]

10 Days of AI Basics, Day 9: Model Deployment, i.e., sending the model into the real world

After collecting data, training your model, and validating its performance, you're ready to share it with customers. You're ready for model "deployment." Unfortunately, it can actually be quite complicated.

What Is Model Deployment?

Deployment is when your model graduates from development to production. It's the transition from "this works on my laptop" to "this works for millions of users." It's similar to the difference between cooking a great meal in your kitchen vs. running a restaurant. The skills are related, but the challenges are vastly different. So what exactly are the challenges with deploying AI models?

1. Scaling

Your model needs to handle *many requests at the same time.* During development, you probably ran inference on one or a few examples and/or users at a time. That's fine for prototyping. In production, however, thousands of people might query your model simultaneously.

You need to think about:
- Hardware: What servers will host your model?
- Load balancing: How do you distribute requests across servers?
- Latency: How fast can you return responses?
- Cost: GPUs are expensive; how do you manage compute costs at scale?

2. Serving

Your model needs to be accessible via code. This usually means setting up an API; other applications can send requests to your model and receive responses programmatically.

Questions to consider:
- Where will you host the model? (Cloud provider? Self-hosted?)
- What's your API design?
- How do you handle authentication and rate limiting?
- What happens when something goes wrong?

3. Monitoring

Your model's performance will change over time. Why? Because the world changes. The data distribution shifts. Edge cases emerge that you never anticipated during training.

You need to:
- Track performance metrics in production
- Set up alerts for when things go wrong
- Collect examples where the model fails
- Plan for periodic retraining

A model isn't a one-and-done thing. It's a living system that needs ongoing attention.

4. Retraining

Eventually, you'll need to update your model. Maybe performance has degraded. Maybe you have new data. Maybe you've discovered edge cases that need to be addressed.

This creates new questions:
- How do you retrain without disrupting service?
- How do you validate the new model before deploying it?
- How do you roll back if something goes wrong?
- How do you version your models?

DEPLOYMENT: The Hidden Complexity

Surprisingly, training a model might actually only make up 20-30% of the work. The rest is everything that comes after: data pipelines, infrastructure, monitoring, retraining, operations. It's called "MLOps" - machine learning operations - and it's become a discipline of its own.

So yes, model deployment can be complicated. That said, perhaps ironically, this is all a lot easier now, thanks to AI, which can guide us through it.

[Olawale Badekale @cuthbert_ola]

Day1/365 👨‍💻

AI/ML

[Godson @This_is_Godson]

The most slept-on side gig for earning in 2026 isn't coding. It’s AI Training.
Not building the models.
Not writing the math.
Teaching them how to think. 🧵👇

[Vaishnavi @_vmlops]

A platform offering remote data annotation work with flexible hours
https://t.co/xbAjjTM4Uc

[Olawale Badekale @cuthbert_ola]

Sometimes, my mama's house don't feel like home
Sometimes, my baby's arms don't feel like home
Sometimes, with my friends, I feel alone
Sometimes, I feel alone

[Olawale Badekale @cuthbert_ola]

Sometimes, my mama's house don't feel like home
Sometimes, my baby's arms don't feel like home
Sometimes, with my friends, I feel alone
Sometimes, I feel alone