Espoir Murhabazi
@esp-py.bsky.social
#MwanaMboka 🇨🇩 trying to survive in 🇬🇧.
Deploying Machine Learning models one pod and yml file at a time.
PS: You don't need a GPU to do ML
#Mancity 🦈
Deploying Machine Learning models one pod and yml file at a time.
PS: You don't need a GPU to do ML
#Mancity 🦈
Your brain is worth billions. Build empires, not apps.
While the world obsesses over the next GPT wrapper, we will forget the next industrial revolution….
Not My quote.
While the world obsesses over the next GPT wrapper, we will forget the next industrial revolution….
Not My quote.
April 1, 2025 at 2:08 PM
Your brain is worth billions. Build empires, not apps.
While the world obsesses over the next GPT wrapper, we will forget the next industrial revolution….
Not My quote.
While the world obsesses over the next GPT wrapper, we will forget the next industrial revolution….
Not My quote.
Reposted by Espoir Murhabazi
You need to recognize when the (bad) code is:
• Overly complex or full of anti-patterns
• Publicly exposing API keys
• Running GPU batch inference via pd.DataFrame.apply()
• Downloading assets with every request (instead of caching)
• Mixing business logic into UI
• Overly complex or full of anti-patterns
• Publicly exposing API keys
• Running GPU batch inference via pd.DataFrame.apply()
• Downloading assets with every request (instead of caching)
• Mixing business logic into UI
March 26, 2025 at 1:25 AM
You need to recognize when the (bad) code is:
• Overly complex or full of anti-patterns
• Publicly exposing API keys
• Running GPU batch inference via pd.DataFrame.apply()
• Downloading assets with every request (instead of caching)
• Mixing business logic into UI
• Overly complex or full of anti-patterns
• Publicly exposing API keys
• Running GPU batch inference via pd.DataFrame.apply()
• Downloading assets with every request (instead of caching)
• Mixing business logic into UI
Reposted by Espoir Murhabazi
Here’s the 2025 LLM roadmap 😊
1. Code and train your own LLM to really understand the fundamentals
2. Train models more conveniently using production-ready libraries
3. Learn about the big-picture considerations for real-world LLM/AI apps
1. Code and train your own LLM to really understand the fundamentals
2. Train models more conveniently using production-ready libraries
3. Learn about the big-picture considerations for real-world LLM/AI apps
February 23, 2025 at 7:41 PM
Here’s the 2025 LLM roadmap 😊
1. Code and train your own LLM to really understand the fundamentals
2. Train models more conveniently using production-ready libraries
3. Learn about the big-picture considerations for real-world LLM/AI apps
1. Code and train your own LLM to really understand the fundamentals
2. Train models more conveniently using production-ready libraries
3. Learn about the big-picture considerations for real-world LLM/AI apps
From Engineer to apprentice Historian.
My recent summary of the first two wars in Congo.
Mwana ya moto ata omelisi ye libele na butu na tongo tango akolamuka akotuna wapi baboti baye.
https://www.murhabazi.com/my-summary-of-the-two-congo-wars
#Congo #M23 #RwandaIsKilling #FelixOut #KabilaDeg...
My recent summary of the first two wars in Congo.
Mwana ya moto ata omelisi ye libele na butu na tongo tango akolamuka akotuna wapi baboti baye.
https://www.murhabazi.com/my-summary-of-the-two-congo-wars
#Congo #M23 #RwandaIsKilling #FelixOut #KabilaDeg...
March 18, 2025 at 11:03 PM
From Engineer to apprentice Historian.
My recent summary of the first two wars in Congo.
Mwana ya moto ata omelisi ye libele na butu na tongo tango akolamuka akotuna wapi baboti baye.
https://www.murhabazi.com/my-summary-of-the-two-congo-wars
#Congo #M23 #RwandaIsKilling #FelixOut #KabilaDeg...
My recent summary of the first two wars in Congo.
Mwana ya moto ata omelisi ye libele na butu na tongo tango akolamuka akotuna wapi baboti baye.
https://www.murhabazi.com/my-summary-of-the-two-congo-wars
#Congo #M23 #RwandaIsKilling #FelixOut #KabilaDeg...
Null n'as le droit d'effacer une page de l'histoire d'un peuple car un peuple sans histoire c'est un corps sans ame! Alain Foka.
This is how I am spending my weekend.
No coding just reading, yes reading about our history!
This is how I am spending my weekend.
No coding just reading, yes reading about our history!
March 18, 2025 at 11:03 PM
Null n'as le droit d'effacer une page de l'histoire d'un peuple car un peuple sans histoire c'est un corps sans ame! Alain Foka.
This is how I am spending my weekend.
No coding just reading, yes reading about our history!
This is how I am spending my weekend.
No coding just reading, yes reading about our history!
Here you go: github.com/espoirMur/co...
github.com
February 7, 2025 at 11:08 PM
Here you go: github.com/espoirMur/co...
March 18, 2025 at 11:03 PM
That was me, giving a talk at @pydatalondon.bsky.social ! More to come..
Finally tonight we’re hearing from Espoir Murhabazi on Building a News Summariser on a Budget.
A fully fledged LLM application using Qwen2-Instruct 1.5B, Scrapy, Stella embeddings model and hierarchical clustering in plain scipy!
A fully fledged LLM application using Qwen2-Instruct 1.5B, Scrapy, Stella embeddings model and hierarchical clustering in plain scipy!
February 5, 2025 at 1:35 PM
That was me, giving a talk at @pydatalondon.bsky.social ! More to come..
Reposted by Espoir Murhabazi
Finally tonight we’re hearing from Espoir Murhabazi on Building a News Summariser on a Budget.
A fully fledged LLM application using Qwen2-Instruct 1.5B, Scrapy, Stella embeddings model and hierarchical clustering in plain scipy!
A fully fledged LLM application using Qwen2-Instruct 1.5B, Scrapy, Stella embeddings model and hierarchical clustering in plain scipy!
February 4, 2025 at 8:43 PM
Finally tonight we’re hearing from Espoir Murhabazi on Building a News Summariser on a Budget.
A fully fledged LLM application using Qwen2-Instruct 1.5B, Scrapy, Stella embeddings model and hierarchical clustering in plain scipy!
A fully fledged LLM application using Qwen2-Instruct 1.5B, Scrapy, Stella embeddings model and hierarchical clustering in plain scipy!
@DeepIndaba While I agree that Rwanda is not the only country to blame in this conflict and that we Congolese should hold our leaders accountable for what is happening I would like to raise my voice as an Indaban and recommend to revaluate the decision to host the 2025 event in Kigali.
March 18, 2025 at 11:03 PM
@DeepIndaba While I agree that Rwanda is not the only country to blame in this conflict and that we Congolese should hold our leaders accountable for what is happening I would like to raise my voice as an Indaban and recommend to revaluate the decision to host the 2025 event in Kigali.
Dear @DeepIndaba Organizers.
The recent events in Congo have proved to the rest of the world that Rwanda similarly to Russia has violated human rights with its support to the M23 rebels.
1/n
The recent events in Congo have proved to the rest of the world that Rwanda similarly to Russia has violated human rights with its support to the M23 rebels.
1/n
March 18, 2025 at 11:03 PM
Dear @DeepIndaba Organizers.
The recent events in Congo have proved to the rest of the world that Rwanda similarly to Russia has violated human rights with its support to the M23 rebels.
1/n
The recent events in Congo have proved to the rest of the world that Rwanda similarly to Russia has violated human rights with its support to the M23 rebels.
1/n
It's sad to hear the news in Congo!
At least a few think we can do is to share what Is happening:
Here is my daily post on Reddit on the general situation as of Today:
https://www.reddit.com/r/Congo/comments/1id63kx/my_summary_of_the_situation_happening_in_goma_drc/
For those who want to hel...
At least a few think we can do is to share what Is happening:
Here is my daily post on Reddit on the general situation as of Today:
https://www.reddit.com/r/Congo/comments/1id63kx/my_summary_of_the_situation_happening_in_goma_drc/
For those who want to hel...
March 18, 2025 at 11:03 PM
It's sad to hear the news in Congo!
At least a few think we can do is to share what Is happening:
Here is my daily post on Reddit on the general situation as of Today:
https://www.reddit.com/r/Congo/comments/1id63kx/my_summary_of_the_situation_happening_in_goma_drc/
For those who want to hel...
At least a few think we can do is to share what Is happening:
Here is my daily post on Reddit on the general situation as of Today:
https://www.reddit.com/r/Congo/comments/1id63kx/my_summary_of_the_situation_happening_in_goma_drc/
For those who want to hel...
If you want to learn science:
I will still recommend Andrew Ng's ML course, ISL, and a Good Software Engineering book.
Make sure you understand :
- Regression, Linear and Logistic
- Tree methods and their extension (XGBoost)
- Clustering Algorithms
- Neural Networks.
I will still recommend Andrew Ng's ML course, ISL, and a Good Software Engineering book.
Make sure you understand :
- Regression, Linear and Logistic
- Tree methods and their extension (XGBoost)
- Clustering Algorithms
- Neural Networks.
March 18, 2025 at 11:03 PM
If you want to learn science:
I will still recommend Andrew Ng's ML course, ISL, and a Good Software Engineering book.
Make sure you understand :
- Regression, Linear and Logistic
- Tree methods and their extension (XGBoost)
- Clustering Algorithms
- Neural Networks.
I will still recommend Andrew Ng's ML course, ISL, and a Good Software Engineering book.
Make sure you understand :
- Regression, Linear and Logistic
- Tree methods and their extension (XGBoost)
- Clustering Algorithms
- Neural Networks.
Reminder:
GenAi will not solve all your data science problems.
The traditional methods still have their value in a data science workflow.
Define your problem and the business value you will solve with data science, get the data, train Xgboost, and write a pipeline and ✅
GenAi will not solve all your data science problems.
The traditional methods still have their value in a data science workflow.
Define your problem and the business value you will solve with data science, get the data, train Xgboost, and write a pipeline and ✅
March 18, 2025 at 11:03 PM
Reminder:
GenAi will not solve all your data science problems.
The traditional methods still have their value in a data science workflow.
Define your problem and the business value you will solve with data science, get the data, train Xgboost, and write a pipeline and ✅
GenAi will not solve all your data science problems.
The traditional methods still have their value in a data science workflow.
Define your problem and the business value you will solve with data science, get the data, train Xgboost, and write a pipeline and ✅
Let us not ban them, because they will always find a way to circumnavigate the ban. Instead, give them the tool and show them how to use them to their advantage to learn.
March 18, 2025 at 11:03 PM
Let us not ban them, because they will always find a way to circumnavigate the ban. Instead, give them the tool and show them how to use them to their advantage to learn.
With the rise of LLM and GenAi, we need to rethink how we teach the young generation.
With the access kids have to those tools today, traditional teaching methods will not work anymore. We should think about how to teach kids to leverage tools like ChatGPT to learn.
With the access kids have to those tools today, traditional teaching methods will not work anymore. We should think about how to teach kids to leverage tools like ChatGPT to learn.
March 18, 2025 at 11:03 PM
With the rise of LLM and GenAi, we need to rethink how we teach the young generation.
With the access kids have to those tools today, traditional teaching methods will not work anymore. We should think about how to teach kids to leverage tools like ChatGPT to learn.
With the access kids have to those tools today, traditional teaching methods will not work anymore. We should think about how to teach kids to leverage tools like ChatGPT to learn.
My daily drive for three years has let me down!
I have been using my Remarkable 1 since 2021.
It served me in my Uni days, and my first days at work.
I don't know what happened to him last week, it went down.
I miss him.
Who should I get next RM2 or Supernote? 🤔
I have been using my Remarkable 1 since 2021.
It served me in my Uni days, and my first days at work.
I don't know what happened to him last week, it went down.
I miss him.
Who should I get next RM2 or Supernote? 🤔
March 18, 2025 at 11:03 PM
My daily drive for three years has let me down!
I have been using my Remarkable 1 since 2021.
It served me in my Uni days, and my first days at work.
I don't know what happened to him last week, it went down.
I miss him.
Who should I get next RM2 or Supernote? 🤔
I have been using my Remarkable 1 since 2021.
It served me in my Uni days, and my first days at work.
I don't know what happened to him last week, it went down.
I miss him.
Who should I get next RM2 or Supernote? 🤔
March 18, 2025 at 11:03 PM
Both call themselves software engineers, and both tend to earn similar salaries in their early careers. But Type 1 engineers progress toward more fulfilling and well-remunerated work over time, whether that’s valuable commercial work breakthrough open-source projects,
2/2
2/2
March 18, 2025 at 11:03 PM
Both call themselves software engineers, and both tend to earn similar salaries in their early careers. But Type 1 engineers progress toward more fulfilling and well-remunerated work over time, whether that’s valuable commercial work breakthrough open-source projects,
2/2
2/2
There are two types of software engineers: those who understand computer science well enough to do challenging, innovative work and those who just get by because they’re familiar with a few high-level tools.
Which one are you?
1/2
Which one are you?
1/2
March 18, 2025 at 11:03 PM
There are two types of software engineers: those who understand computer science well enough to do challenging, innovative work and those who just get by because they’re familiar with a few high-level tools.
Which one are you?
1/2
Which one are you?
1/2
Unpopular opinion…
Next game we start Halaand on the bench…
#MCILIV
Next game we start Halaand on the bench…
#MCILIV
March 18, 2025 at 11:03 PM
Unpopular opinion…
Next game we start Halaand on the bench…
#MCILIV
Next game we start Halaand on the bench…
#MCILIV
I decided to be quiet now because I have a wife to take care of.
March 18, 2025 at 11:03 PM
I decided to be quiet now because I have a wife to take care of.
The day, I am drunk enough to write a post about my life in the UK there will be fire...🔥
Over the past three years, I have learned a lot about capitalism and the corporate world...
The day I write a post, the fire will be bigger than what got me fired at Andela.
Over the past three years, I have learned a lot about capitalism and the corporate world...
The day I write a post, the fire will be bigger than what got me fired at Andela.
March 18, 2025 at 11:03 PM
The day, I am drunk enough to write a post about my life in the UK there will be fire...🔥
Over the past three years, I have learned a lot about capitalism and the corporate world...
The day I write a post, the fire will be bigger than what got me fired at Andela.
Over the past three years, I have learned a lot about capitalism and the corporate world...
The day I write a post, the fire will be bigger than what got me fired at Andela.
Tous ceux qui sont a la recherche de l'amour dans les relations internationales ficher le camp.. Donald #Trump ne nous aime pas, #Macron ne nous aime pas, Poutine ne nous aime pas! C'est normal ce ne sont pas les africains ce n'est pas leur role de nous aimer!
Dixit Nathalie Yamb
Dixit Nathalie Yamb
March 18, 2025 at 11:03 PM
Tous ceux qui sont a la recherche de l'amour dans les relations internationales ficher le camp.. Donald #Trump ne nous aime pas, #Macron ne nous aime pas, Poutine ne nous aime pas! C'est normal ce ne sont pas les africains ce n'est pas leur role de nous aimer!
Dixit Nathalie Yamb
Dixit Nathalie Yamb
3 years to go .......................
March 18, 2025 at 11:03 PM
3 years to go .......................