BUILD DEEP LEARNING PRODUCTS, attract more opportunities

Step 1: Extend your “ML Product Dev” knowledge

Step 2: Build one for a real-world problem

Step 3: Win Cash Prize


Techniques: AI Product Development, MLOps and Engineering, Natural Language Processing, Knowledge Graphs, Graph Neural Networks, Document Classification, Explainable AI


Deep Learning Product Challenge

Step 1: Learn how to build an ML Product

Step 2: Build one for a real-world problem

Step 3: Win Cash Prize


Techniques: AI Product Development, MLOps and Engineering, Natural Language Processing, Knowledge Graphs, Graph Neural Networks, Document Classification, Explainable AI



Attract more opportunities by strengthening and diversifying your portfolio by showing your experience in building ML products for real life problems


Solve interesting business problems (your own or a sponsored project idea) in a structured and guided experience


Work with a team of 2-3 peers to solve a business challenge using AI and win awesome prize money when your MVP is selected as the best solution!


Get guidance from senior industry practitioners on ML research, engineering, and product development


Access our entire library of expert curated content to quickly learn whatever you need to solve your problem


Learn from peers in teams of 2-3 throughout the program. Connect with like-minded individuals in our 10,000+ community of data scientists and ML engineers

Potential Use Cases & Cash Prizes

Sponsor Company: will be disclosed after registration
Cash Prize: $2,500
Problem Statement: Given a news article, identify whether it points to a potential financial risk for any of the companies mentioned in it
Context: Tracking financial health of a company is a crucial problem to solve for anyone that intends to invest in the company. Although structured data – which includes share prices, trading volumes, etc. – can help in building mathematical models that can aid in solving this problem, tapping into unstructured data such as news articles about the company may prove to be useful in identifying certain subtle trends, events which may lead to deterioration of company’s financial health. Scalable models that can identify news items that could potentially point to a financial risk, from a large volume of news items are thus useful when making investment decisions.
Objective: A good solution should be able to handle large volumes of inputs at inference time and process them in reasonable amounts of time. From a bulk of news articles, the solution should be able to infer the articles as risk/non-risk, with highest precision and recall achievable.
Dataset: Sponsor company will provide a starting dataset. However, participants are encouraged to explore other possible datasets that might help them improve their pipeline performance.
Techniques Used: Natural Language Processing, Document Classification, Entity Recognition, Information Extraction 
Sponsor Company: Aggregate Intellect
Cash Prize: $1,000
Problem Statement: How might we automatically curate an ordered list of resources relevant to a particular use case?
Context: Users who are interested in learning a particular topic quickly can either spend a lot of time searching, or can take courses. The issue with searching is that it’s open ended and won’t provide any particular structure to the user to follow. The issue with courses is that they are pre-determined and hard to relate directly to any particular use case user has.
Objective: Given a user query, shortlist 10 articles that are most relevant, and order them based on which article is prerequisite for the next one. in other words, instead of just ranking the results based on semantic relevance to query, we want to rank a limited number of them based on how they relate to each other from a concept dependence point of view.
Dataset: Sponsor company will suggest some open source datasets. However, participants are encouraged to explore other possible datasets that might help them improve their pipeline performance.
Techniques Used: Natural Language Processing, Knowledge Graph, Graph Embeddings, Recommender Systems

Have you been wanting to kick-start a project for that venture you’ve always dreamt of?

Maybe you’re hoping to add a shiny new project to your portfolio?

This is where you can start! Join us, team up with like-minded people, build that project.


Enjoy 2 x 1 hour lectures plus Q&A with renowned University of Alberta and Amii professor Matthew Taylor.

Topics are: Reinforcement Learning in the Real World (March 22nd, 12:30pm ET) and Human-Machine Learning Systems (March 29th, 12:30pm ET)


Build an AI product, potentially win a cash prize, and importantly expand and diversify your experience of working on real business problems


Get guidance from senior industry practitioners (project leads) throughout this project to ensure you get the most out of this experience 

More Info

The MasterClass with Matthew Taylor will involve 2 one hour lectures with plenty of time for Q&A!

Lecture 1: Reinforcement Learning in the Real World

March 22nd, 12:30pm ET

Reinforcement learning (RL) has had many successes, from AlphaGo to StarCraft 2 and data center cooling. Unfortunately, many real-world business tasks are not easily defined and can be difficult to formulate so that RL can successfully solve them. Furthermore significant amounts of time and/or data may be required to reach acceptable performance, potentially making learning infeasibly slow or expensive. This talk will: 

  1. Briefly remind the audience what RL is;
  2. Discuss how to effectively identify and formulate an RL problem, including concrete use cases and;
  3. Identify ways to leverage existing knowledge from other agents, programs, or people to help RL agents to improve their performance.

Lecture 2: Centaurs and Augmented Intelligence: Human-AI Systems in 2021

March 29th, 12:30pm ET

Would you want a computer to decide whether you needed an ambulance when you called 911? For the foreseeable future, machine learning and artificial intelligence systems will need to incorporate humans in high-stakes environments. Furthermore, even if your end goal is a completely autonomous solution, it may be much easier to have an initial deployment where the AI is assisting a human. This talk will discuss 

  1. When and why human-AI systems can outperform human-only or AI-only approaches by leveraging the unique abilities of both humans and AIs, 
  2. State of the art methods for successful combinations, and 
  3. Showcase successful examples to help you identify potential centaur approaches in your business.

Matthew E. Taylor received his doctorate from the University of Texas at Austin in the summer of 2008, supervised by Peter Stone. He then completed a two-year postdoctoral research position at the University of Southern California with Milind Tambe and spent 2.5 years as an assistant professor at Lafayette College. He was then an assistant professor at Washington State University where he held the Allred Distinguished Professorship in Artificial Intelligence. In 2017, he temporarily left academia to help start an artificial intelligence lab in Edmonton, Alberta, with Borealis AI, the artificial intelligence research lab for the Royal Bank of Canada, where, among other things, he contributed to a deployed reinforcement learning agent for stock trading.

He is now a tenured associate professor in computer science at the University of Alberta, a Fellow-in-Residence at the Alberta Machine Intelligence Institute, and remains an adjunct professor at Washington State University. He has been a PI or co-PI on over $7M USD in competitively awarded research funding from federal, state, and industrial sources, including a National Science Foundation CAREER award and a Canada CIFAR AI Chair. He has (co-)supervised 7 graduated PhD students and 5 MS students in the Intelligent Robot Learning Lab, as well as published over 120 peer-reviewed papers in conferences and journals. His current fundamental and applied research interests are in reinforcement learning, human-in-the-loop AI, multi-agent systems, and robotics.

Familiarity with the following is expected:
  • Python: pandas, scikit-learn
  • (Optional) PyTorch
  • Jupyter Notebook
  • Stats, ML & Neural Nets
“Do I need a good laptop for this course?”
  • No. The hands-on notebooks are provided in Google Colab
  • Google Colab is a zero-configuration environment with free access to cloud GPU’s. Just make sure you have the latest chrome or firefox browser installed.
  • Recommendations for extra resources are available for your capstone

Need clarification about the prerequisites? 
Join our Slack Community!

You can then either post your questions in #community-questions or DM Amir Feizpour 

  • Data Scientists
  • Machine Learning Engineers
  • Product Managers
  • Software Engineers

Questions / not sure if you make the cut? 
Join our Slack Community!

You can then either post your questions in #community-questions or DM Amir Feizpour 

  • Guided Capstone Competition Begins: Mon, Mar 1st
  • Last Day to Register: Friday, Mar 5th
  • Last Day for Refunds: Friday, Mar 5th
  • Capstone Submission Deadline: Sunday, April 11th



Not in the mood for a competition? No sweat! Take it at your own pace.
$ 39
  • Try out our subscription for a few months
  • Self-paced expert curated content
  • Dozens of Learning Packages
  • Hundreds of videos and shortlists
  • Engaged & welcoming community of experts


Work on a sponsored project and compete for cash prize
$ 99
  • FREE for existing paying subscribers
  • 6-week structured & guided application building experience
  • Weekly group touchpoints & deliverables
  • Collaborative learning with teams of 3
  • Publish results as a short video, a github repo, and a working web app
  • Includes everything in "Learn" package


Want to build something of your own?
$ 339
  • Build an application for your portfolio
  • 2 x 1 hour MasterClass
  • 6 x 20-min 1-on-1's with Experts
  • Includes everything in "Compete" package

Apply for 99% off discount!

There are seats available for anyone who is interested to work on the sponsored projects and has prior experience of handling use cases like those. The number of available seats is limited and the selection process will be competitive, so please make sure you provide as much information as you can.


Week 1
  • Join the Slack Community & Meet your Cohort
  • Watch the AI Product Development Video series (83 minutes)
  • Do market / user research to understand your problem statement better
Week 2
  • Watch MLOps and Engineering Video Series (34 minutes) and do hands-on exercises 
  • Finalize your datasets, do EDA, build base model
  • Get feedback from the cohort and your project lead
Week 3
  • Meet your team (we do team matching based on your explicit preferences + a few other factors)
  • Combine / refine your problem statement and solution as a team
  • Do an alpha release based on everything you learned so far
Week 4
  • Continue getting feedback from your users, project lead, and sponsor company representative
  • Watch / use our advanced ML content (access to 11+ learning packages!) to improve your pipeline based on the feedback
Week 5
  • Get more feedback from your Cohort & dedicated project lead
  • Improve your pipeline
  • Get ready for your final demo
Week 6
  • Clean up / finalize your code and demo video
  • Submit your entry for the competition!

Past Capstones


You're in Good Company

Our community and partners attract an incredible range of AI & ML practitioners looking to stay ahead of the curve. Below are a few selected companies / universities our users represent. 


The answer to this question is almost always "YES"!

In the checkout process, there's a tab called "Discounts" where you can see all available discounts at the time of your purchase:

  • Personalized Discount (0-90%): This is a discount for all our existing users. It is proportional to your recent level of participation and engagement on our platform (using content, participating in events, etc). 
  • Referral Discount (20%): This is discount you get from referring others to our platform where they purchase and we give you 20% of their purchase value as credit
  • Friend Discount (20%): This is a discount you get when a friend refers you to our platform
  • Partner Discount (5-50%): We normally partner with various communities and orgs where they provide extra discount to their members. A good way to find these is to search our name "aggregate intellect" on social media
  • Sponsored Capstone (99%): This is a discount available for company sponsored capstone projects. This is limited and only available through a competitive application process. Apply Here!!!

At the moment our sponsored competitions are only open to the residents of Canada, United States, and Europe. We are working with our lawyers to expand this. 

If you are from other areas, you could still participate in independent capstone build or other paid or free activities.

Not particularly! Obviously, the more you know the better. But, part of our philosophy is "learning by doing" (ehm, I mean "experiential learning" - to be fancier than necessary). We provide content, and expert guidance, and you will be learning a #$#$h load from your peers as you try to build things together. All we need you to bring is your ambition, curiosity, and well, enough coding skills.

Not particularly! Obviously, the more you know the better. But, part of our philosophy is "learning by doing" (ehm, I mean "experiential learning" - to be fancier than necessary). We provide content, and expert guidance, and you will be learning a #$#$h load from your peers as you try to build things together. All we need you to bring is your ambition, curiosity, and well, enough coding skills.

You can modify your purchase until the refund deadline (4 days after the start date, typically). Note, however, that any discounts that have expiry date wont necessarily be available when you modify your purchase. In order to modify your purchase, go through the same process as your did the first time, and in the "customize your purchase" step, add anything you want.

For refunding any existing purchases contact us at [email protected] 

I mean, we can't really stop you if that's what you want! BUT do we encourage it? Not really!!! There is *SO* much learning in doing things together with peers. However, we understand that some folks prefer that. So, just let us know and we will make it happen the way you like it.

The Prizes will be awarded by a panel of judges consisting of representatives of the Sponsor (50%) and our community experts (50%) based on the following:

  1. Effort put in (problem discovery, data collection, training, etc)
  2. Readiness of the solution (can it be used locally or non-locally by many people? how much work is needed for the solution to become minimally viable?)
  3. Optimization (how quickly can your solution perform inference? Points may also be awarded for training optimizations, such as transfer learning.)
  4. Interpretability (how easy is it to explain why your solution works?)
  5. Performance (how well does your solution perform in terms of accuracy - may be a combination of objective and subjective metrics, or other unsupervised performance metrics).
  6. Practicality of the application (is what you built solving an interesting practical problem in a way that can be used by intended users?)
  7. Technical soundness (are you using the right evaluation metrics? Are you using the right algos? Are you interpreting your results correctly?)
  8. Originality & Creativity (were you able to reframe the problem, or perhaps collect / synthesize data, in a creative manner which sidesteps some of the more challenging obstacles? Did you have an original approach to solving the problem or sufficed to overly done approaches?)
  9. Clarity of communication (is it easy to follow your story, and technical detail in your deck and video?)
  10. Quality of production (is the audio/video quality of your final deliverable acceptable) 
  11. Commercial Use (does your solution use copyright or proprietary material, or is it available with an unrestricted license?)
  12. Code Quality (is your code easy for other engineers to understand and modify?)

We estimate around 10 hours a week, but it really depends on your base level, and how much effort you want to put in. It also depends on how effectively you "divide and conquer" with your team members.

The subscription gives you blanket access to a range of premium material (videos, hands on code samples, etc) on topics like

  • ML Product Development
  • MLOps and Engineering
  • Modern Natural Language Processing
  • Reinforcement Learning
  • Recommender Systems
  • and many more

You can keep the subscription as long as you want or cancel it whenever.

See full catalogue here

Well! You're at the right place! We got you covered!! You have 3 options:

  1.  Community: Join our community slack workspace, just hang out there and use all our free content until you're ready. Post any questions you have in #community-questions (or DM Amir Feizpour).
  2. Monthly Subscription:  Purchase the "learn" package above, and get blanket access to dozens of premium expert-curated learning package. Consume at your own pace (and risk)!
  3. Build (without competing): Purchase the "build & compete" package above, ignore the 'compete' part, and just work on a non-sponsored project you and a few peers cook up!

If you are working on a company sponsored project, you would be pre-assigning all intellectual property to them. If your solution is not chosen as a winner, you can still publish your work publicly for your portfolio as long as you respect all confidentiality and data privacy terms you agree to as part of the competition. This concretely means that if you don't win, you can remove all references to the sponsor company (their name & brand, any results obtained from their confidential data, all original data they provided and data you derived from their data, any specific details about their use case, etc) and then publish your work publicly.

A lot of our users join our cohorts to add something to their personal / professional portfolio, or refine their ideas for that awesome venture they've been dreaming about. Since we really like ambitious people like you, aggregate intellect doesn't have any claims over the IP you generate as part of our cohorts. It is all yours! all of it!!! 

Stay in Touch!

Not ready for this but want to stay in touch?

Join our newsletter to stay up-to-date with our upcoming social and partner events + get updated resources and fresh AI news insights right to your inbox every Tuesday!