MY EXPERIENCE AT AMMI

10 min readAug 15, 2020

A Shared Story

My name is Daniel Ajisafe, a beneficiary of the most prestigious Machine Learning program in Africa, The African Masters in Machine Intelligence (AMMI).

My offer from AMMI followed with tremendous opportunities. But as it was my first machine learning (ML) program, I had a couple of burgeoning questions like what’s the program going to be like? How do I prepare? What can I do to maximize my learning in the program?

I could not answer some of these questions and had to pick up a couple of strategies along the way. These strategies helped me and I felt I could have done better if I knew them way earlier. So I have decided to share these little strategies and my experiences, with the hope of helping future AMMI cohorts of students (who are my intended audience) as well as potential applicants, on how they can prepare, maximize their learning, benefit most from the program, and keep up with the wonderful intensity that the program demands.

Please sit back as I would be putting the best of a year into this article.

OUTLINE

The WHY, WHAT and HOW of the AMMI program
The Game Plan
***Orientation and Onboarding
***Application deadlines → Interview Season
***Working in Teams
***Supervisor Request Emails
***Final Project
General Advice

Why AMMI program?

For decades, Africans have mostly been consumers of technology made by other counterparts in the world. In this 21st industrial revolution, Artificial Intelligence (AI) is fast becoming the new electricity and there is a need to put ourselves as one of the frontiers of this new technology in order to solve our own problems in our own and unique way. This need became a calling and our director Moustapha Cisse decided to take this call.

How does the AMMI program work?

The AMMI program is an intensive 1-year master’s program for African applicants.

It teaches its students technical concepts from the basics of programming to advanced topics in AI. Its major arm includes structured coursework, labs, practicals, projects, and exam assessment. My coursework started with an AI Bootcamp followed by a total of 10–12 courses.

Unlike conventional school models, where students have to write all their exams in like a month after taking all coursework, the AMMI program validates understanding the concepts for each course through practicals, projects, sometimes group presentations, and exam assessment at the end of the course.

The Game Plan

One of the keys to success in any graduate school is the ability and willingness to be proactive, which means taking responsibility for your own experiences. Pursuing a graduate degree in Machine Learning is an important investment in your personal growth and future career. [1]

The Game Plan provides some form of advice on strategies I think would help to gain the most value from your investment [1]. These steps, however, are my opinions based on my experience in the program.

Before you arrive: Try to do some proactive planning, and planning begins with self-awareness. [1] You are going to be learning in the same classroom with the best of the best students in Africa, with one common goal which is IMPACT. To do this you need some skills, and that’s what the AMMI program is going to build in you. But how do you prepare yourself to learn these skills?

I have found the following courses to be beneficial and serve as a strong pre-arrival foundation for the classes you will take during the program. Some of them would also prepare you for what I call the “Interview Season”.

Mathematics for Machine Learning Specialization (Coursera): This course is gold for students who do not have a strong mathematical background or would like to refresh their mathematical prowess. Don’t forget AI sits on Maths, Logic, and Programming. Coursera offers lots of free courses but for this one, you will have to apply for financial aid and most people get it (I did 😄). Feedback for the financial aid takes a total of 2 weeks so it is good to apply early.
Introduction to Data Analysis: This free course on Udacity introduces you to the world of data analysis. You will learn how to use Python libraries like NumPy, Pandas, and Matplotlib to write code that’s cleaner, more concise, and runs faster.
Data Structures and Algorithms: This course is a gem for the Interview season. It’s free on Udacity. It explains common data structures and algorithms in Python for technical interviews.
Familiarize yourself with Jupyter Notebook as a working environment and be comfortable using the Linux operating system.

Now that you have understood some pre-arrival foundation, let’s get into the program. Welcome to AMMI!

The Ammi Roadmap

I) Orientation & Onboarding

Acceptance to the AMMI program includes a full scholarship, thanks to wonderful sponsorship from Facebook and Google, to cover for your housing, meals, transportation, health insurance, and a monthly stipend.

The program begins almost as soon as you arrive, meaning you won’t have to worry about any of the above. We had our orientation in the first week and a city tour courtesy of the school. 😄

II) Application Deadlines

October to Late December is the application season for a lot of research programs. AMMI provides a good platform to get into these programs. These include AI Residencies, research Internships, MSc/Ph.D. positions in Machine Learning, Applied AI positions, etc.

Please be prepared, the applications can be overwhelming. Don’t forget you are going to be doing this in tandem with AMMI lectures, working on assignments, and meeting the school’s deadlines. I have created a repository with some of the opportunities I have applied for during the program. This includes residency programs, research internships, etc.

The AMMI program teaches and requires its students to implement concepts that underline the understanding that most of these AI powerhouses ask for during interviews. I will share some general AI interview steps and focus only on the residency programs, leading up to the interview season. I believe this could be generalized for other interviews including internships. I have included a separate link to the interview season here.

III) Working in Teams

There is a popular saying that “If you want to go fast, go alone. If you want to go far, go together”.

I am a continual beneficiary of these saying. You can’t come to AMMI and expect to work in isolation. Teamwork is very important and a core practice if you wish to succeed during the program.

The best student in the AMMI program is not the smartest student in the class, but the best student is he who would strive to bring out the best in others. For this reason, you could perhaps find the AMMI program a little different from conventional schools because we thrive on togetherness, team spirit, impact, and giving back.

IV) Supervisor Request Emails

As you get somewhere toward the middle of the program, if you wish to work in a specific AI field or with a specific supervisor, you would need to start what I call the “Supervisor Hunt” 😅
No sweats! There are a lot to choose from but to do this effectively, you need to be proactive.

No 1. To have someone supervise you requires that you demonstrate an area of interest that aligns with the potential supervisor. However, it is possible that you don’t have a particular area of interest at this time in the program but it’s generally advisable to find your interest early. At this stage, the program would have introduced you to core topics in AI enough to help you make this decision.

AMMI has a network of wonderful supervisors who would be happy to take you on but you will need to work on your presentation and initial correspondence.

I have a draft for this and I hope it helps.

Dear (Dr.xxx) / xx xxx,I hope this email finds you well. My name is xxxxx, a student in the current cohort of the African Masters in Machine Intelligence (AMMI) program.AMMI is a one-year master's degree program hosted by the African Institute for Mathematical Sciences (AIMS) with sponsorship from Google and Facebook.Towards the end of the AMMI program, we are required to start working on a final project to fulfill the requirement for the master's degree partially.I am, for these reasons, reaching out to you to kindly consider mentoring and possibly advising me on my final project. I am broadly interested in XXXXXX. I am also very excited about exploring and trying new research ideas that would potentially increase my learning curve.I share your research interests in YYYYY and find your publications very interesting. One I recently read was "ZZZZZ" where you worked directly on panoramic images, rather than decomposing them into perspective images as other approaches do.I have attached my CV for your perusal, and I am willing to discuss further, should you be interested in my request. I'll be happy to hear from you.Thank you in anticipation for your kind consideration and acceptance.

DO NOT copy-paste the above.. LOL! Make sure to understand from the above the type of content that supervisors would appreciate reading from you. Needless to say, I imagine that by the time you reach out to your supervisor, you would have read her/his work.

You are a student in one of the best ML programs in the world, fully funded by Google and Facebook. Please flaunt this 😆 on your resume, motivation letter, and emails. Oh, and students from previous cohorts have interned at Google, Facebook, Vector Institute, MILA, and many more amazing places. It is a once in a lifetime opportunity and I wish you to treat it as such :)

No 2. Widen your net: It is possible the research scientist or Professor you have always admired might not be available to supervise you. This is fine. There are lots of other researchers in your area of interest who would be happy and available to supervise you. Hence, widen your net or better still, increase your hyper-parameter space 😃

V) Final Research Project

This is a final capstone that demonstrates what you have learned so far. The program encourages students to start working on their projects as soon as the coursework part of the program nears completion, somewhere in June/July. You would need to accomplish IV to get to V.

The outcome of the project is a technical report/writing that follows the format of a NeurIPS style paper with a maximum of 10 pages. Here is a sample report by alumni Kobby Panford-Quainoo, which got accepted at the PML4DC workshop for ICLR 2020.

General Advice

Now that we have come to the end of our roadmap, I would like to share some general advice that helped me during the program.

Faith: I am a believer, so is also my Christian and Muslim brothers in the program. Acknowledging God as my most sure foundation gave me strength in days when I got tired. His acknowledgment cannot be over-emphasized because everything comes and leads back to Him. Thank you Jesus!
Taking breaks: Sometimes I can be boxed up trying to solve a problem or understanding a new concept. To maintain sanity, I take a walk, watch a movie, or go to sleep. Often times, the answers we seek can be found in places we don’t expect. Ask Albert Einstein, he was led to discover the law of gravity when he took a nap under the tree!

3. Giving back: After experiencing the AMMI program, you will most likely carry on the culture of giving back (i.e tutoring a student or more from your previous school, mentoring the next AMMI cohorts, joining the outstanding group of resident tutors for the program, etc.)

4. Meet People & Learn about your new environment: My host country was Rwanda. My colleagues visited Lake Kivu (I did not, so sad!). We went to Kigali Arena, watched the National BasketBall team play a final game, and visited the Museum for Campaign against Genocide.

Plus I learned a few Rwanda words and could communicate with the native people (or at least take a motorbike with it 😄). Please try new food and also explore the city!

My AMMI program was a life-transforming experience, and it has set me on the right path to achieving my goals. The world is at your feet and the ball is in your court. Take the initiative now, be proactive, and start learning!

I would like to say a big thank you to my colleagues and amazing mentors for the wonderful feedback I received.

Again, I say WELCOME TO AMMI!