Tiny Machine Learning on HarvardX Q&A / AMA on March 18th @ 430pm EST


On Thursday, March 18th, at 430pm EST, we are holding our first AMA (ask me anything) session.

About the Event

You will have the opportunity to meet the instructors (Laurence, Pete and Vijay) and the staff behind TinyMLx (Ben, Brian, Colby, Dhilan, Matthew and others) to ask questions. The AMA session is open to all learners from the TinyML HarvardX series, as well as the broader ecosystem. But a gentle request that we’d like to help our learners, first, so please allow this sign-up for our online learners who we are trying to nurture and support.

To all our learners, we want to connect with you to learn about your unique experiences and motivations for taking TinyML. Also, this is an opportunity for you to ask us anything: about the courses, tinyML career pathways, projects, reasons for doing the course the way we did it, etc. We want to support your curiosity, foster your growth, as well as learn from your experiences! We recognize learners are geographically distributed, so this will be the first of many if learners are interested. So be patient if we fill-up this event.

When & Where

March 18th, online via Zoom.

If you are interested, please register by clicking the link below.

The event is (tentatively) limited to 300 attendees!

Q&A / Ask Us Anything

We are collecting questions ahead of time that we will curate to answer live, so please post your question(s) below by replying to this thread. We look forward to your questions & meeting you soon!

VJ & Staff.


Some of the topics we plan to cover include:

  • Course 4: Advanced TinyML, coming soon for our learners! Find out more about it.
  • Projects: Ideas, guidelines and suggestions for our Course 3 learners. More about this.
  • TinyML for STEM: Open courseware for educators who want to teach TinyML on their own.

(and more)

Got questions about any of these, or something else? Please feel free to jot them down here. We will answer them.


[this is not a question for QNA] @vjreddi I would like to join the session! However, 430pm EST is midnight for me here. Can you record the session and share it here?

1 Like

I’d like to deploy a TinyML that detects a visitor my yard - bird, squirrel, insect. And then triggers a digital camera to take a high resolution picture.

What kind of digital cameras would be suitable for this?


Hey @Albert, for sure! We will record and release.

I recognize the time difference is an issue. That’s why this could be the first of many in the future (if there is demand) so that we accommodate all learners by hosting future events at different times. :slight_smile:


Thank you very much for hosting this event, what a fantastic opportunity this is! I am looking forward to getting to meet/learn more about all of you who have made this course happen.

I have not finished Course 3 yet so some of my questions may be answered already, but I’d like to learn more about:

  1. TinyML career pathways - what would you recommend to someone who may want to pursue TinyML professionally?

  2. the bleeding edge of TinyML - As people at the forefront of the industry, what do you think will happen in the next 1-3 years in the world of TinyML?

  3. best practices for explaining TinyML - How would you explain what we are working on to people who may have limited technology experience/interest?

  4. a day in the life of TinyML - What do those of you who are working in the field now consider an average day?

Any insight in to any of these would be great. Looking forward to seeing what else is covered!


These are awesome questions, thanks @stephenbaileymagic for jotting them down. Will make sure we get to these.


Suppose if there are larger models which dont fit into our flash memory, is it a realistic idea to store them on the native SD card (just like in a teensy 4.1)and run the inference using dma transfers?

1 Like

This event is very interesting and much appreciated as well as my journey to the three courses so far.
As an engineer and educator I’m very interested to the TinyML for STEM topic. I would like to get an idea on how to bring TinyML projects to my classroom and lab. What do you believe the prerequisites should be for my students to be able to enter this amazing world?

Looking forward for the event!

1 Like

Hi @gkapsid, welcome! Could you please give us some feedback on what age groups you are thinking about here for STEM education? We’ve got folks who are interested in K-12 AI education as well as college level.

@brian_plancher and @dhilanr will also be on the call and they are helping cook the TinyML4STEM program. Would be great if a few of us all got together to help curate the requirements.

This is a quick questions, what is the name and type of the connectors to use on the TinyML shield in order to use other type of sensors. I know the shield has six ports, but what are the suggested connectors to use, Also the Tiny ML courses are great !!! thank you for creating this type of content much appreciated.

Marlon Ortiz

1 Like

My students are mostly adults which are back to school to improve their professional status. Last year after graduation a pair of them started a start up company on IoT area which was completely unknown to them before. They are usually skilled workers and professionals and a few of them have the potentials to change their career path.
This year though I changed my teaching position and I interact with students that will become primary school teachers (elementary school teachers) ! There I cooperate with professors that teach the use of new technologies (ICT) in classroom for kids between 6-12 yo. AI and how to use it in classroom is under consideration and of great interest.
To conclude I personally have a strong interest in bringing new edge technologies to classroom either as teaching subjects or as part of projects implemented given other opportunities (school contests, robotics clubs, teams etc).
The original question may become: What is prerequisite for a student that has an electronics background and and maybe some experience as a professional?
What should be necessary for a kid 10-14 yo to enter the AI and ML or even TinyML world the way it happened the previous years with robotics (like Lego systems or Arduino clones)?
I would be glad to hear some ideas and of course to contribute in any possible way!

1 Like

If your students had basic programming I think they could dive right in right now (and at least take away a high level understanding of the topics in the edX course).

I do think there is definitely room for some more content at much more introductory level that either abstracts away a lot of the code and/or explains basic coding through the lens of TinyML since both arduino and python are quite friendly! I think we’d need to create / adapt some of the material but/and I think that given all the success of other Arduino based programs for younger students this should be totally possible!

1 Like

These are exactly my thoughts too. Software abstraction is a key element but I believe that for TinyML applications a kind of hardware abstraction is needed too. Fortunately Arduino and other manufacturers have added a few (or more) sensors on board which removes the necessity for complicated connections. The kit you designed for course 3, I find that is also in the right direction.
Do you believe that a Scratch like environment that will allow the development and deployment of simple TinyML applications will show up soon?
To me another key point is the training of the trainers which would be nice to be supported by the technical knowledge as well as lesson plans and examples on how to incorporate TinyML in classroom.
So I believe that

  1. Abstraction of code (maybe using code bricks )
  2. Easy to use hardware (in the form of a kit with no or a few easy connections)
  3. Examples in the form of teaching material
    are a good start.

Yah agree completely! This makes total sense. I think the TinyML Kit with the shield does make the hardware part mostly solved. We are working on uploading the slides / readings / colabs to github which can be a good start to the teaching materials (although definitely needs to be a bit more packaged / revised). The only real sticking point is the abstraction of code which I think would require some careful work to be done correctly (slash with bricks that would be general enough to be worth playing with for students).

1 Like

Thank to all the staff for these very interesting courses. :grinning: :laughing: I learn a lot.
Career pathways is the topic which interest me the most. I am a medical student with programing background. And I’m now very interested to enter in the field of eHealth/Digital health/BME (mostly in ML field) .
I read some documents about this kind of profile (medicals doctors which work in the tech field in company like Google,… and even creating their own startup/company),
but I would like to have some direct advice from Technology Professionals that you are.
Maybe can help someone else in the case too.

Thank you very much!!

I’m Amos from Nigeria. Thanks to the EdX and HarvardX team for making this course available. I’ve registered for the AMA session. I would like to know if for the third module (TinyML deployment) I can use Arduino UNO R3 instead of Arduino Nano BLE sense. That’s the one I use and I wonder if it can be useful in the TinyML deployment.
I look forward to your response. Thank you.

Unfortunately, no. The course requires the use of a variety of sensors that are on the board that aren’t there on the UNO board. It is possible to add sensors but that complicates issues here.

Alright :slight_smile:
Thanks for your response. I’ll have prepared for that. Is the delivery of the product available in Nigeria?

Hi @Amos_Lee,

Yes, the kit should be available. You can order it here:


Please let me know if you have an issue for some reason.