In this talk, I'll journey from my time as a Research Assistant at the Bernoulli Institute, delving into the classification of neurodegenerative diseases, to my encounters with groundbreaking biotechnology and AI companies like Proteinea, AlProtein, Rology, and Natrify in Egypt. These innovative ventures are reshaping industries from their Egyptian hub. Join me as I illuminate the transformative power of this thriving ecosystem, showcasing Egypt's remarkable strides in biotech and AI on the global stage.
5CL-ADBA,5cladba, Chinese supplier, safety is guaranteed
[DSC MENA 24] Medhat_Kandil - Empowering Egypt's AI & Biotechnology Scenes.pdf
1. Empowering Egypt's AI
& Biotechnology Scene
Catalyzing Innovation at the
Intersection of AI and Biotechnology
Presented by Medhat Kandil
2. About me
Research Assistant, University of Groningen | NL
Started my career at a university in the Netherlands, focusing on
neurodegenerative disease classification.
Data & Business Analysis, Sequence Ventures
Spent a year at this Egyptian deep tech VC, gaining exposure to the
local startup ecosystem.
Data Analysis, Contentsquare
Worked in data analysis at this exceptional unicorn startup.
Machine Learning Eng, Gulf Researcher
Currently part of their new Data & AI department, driving an AI-first
approach. Specializing in state-of-the-art Generative AI technology.
Medhat Kandil
Machine Learning Engineer,
Gulf Researcher
3. Studies
Our Focus?
Our Mission?
My Journey In Research
During my Bachelor’s Degree in the Netherlands, I had the
privilege of working as a research assistant at the Bernoulli
Institute and the University Medical Center of Groningen.
Analyzing FDG-PET scan measurements from 127 patients. These
scans reveal neural injury distribution in the brain
Develop accurate supervised learning models to predict
neurodegenerative diseases
The Data?
The dataset included one million voxels per row, categorized into
four classes: Alzheimer’s Disease, Parkinson’s Disease, Dementia,
and Healthy Control.
4. My Thesis
My Goal?
Then What?
Passion drove me to integrate this research into my thesis. I
delved into dimension reduction and normalization techniques.
My goal was to find the optimal dimensionality reduction &
normalization techniques to achieve high accuracy and F1-score,
with a focus on recall.
Our Findings?
We were able to get a reconstruction error of just 7% on the 58
most critical features while applying PCA. This was a very good
result with comparison to the initial one-million feature per
datapoint.
5. Working with
this team
Drove me to
the search for
pioneers in
the region
Ignited my
passion for AI
applications
in biotech
Made me
determined to
drive impact
in my country
and region
6. My time at Sequence Ventures
My interest in deep-tech startups and the
dynamics of the venture world led me to
Sequence Ventures.
In my time there I have gained exposure to
Egyptian startups in AI, Biotechnology,
Blockchain, and the intersections of these
industries.
From there, I have developed a focus on the
companies leveraging AI in Biotechnology.
7. Providing teleradiology
services in African countries
using AI.
Machine Learning methods
assisting teleradiologists in
annotating and diagnosing
scans.
Impact on healthcare
accessibility in underserved
communities.
Rology
The 3 Musketeers of AI &
Biotechnology
Using AI to produce
alternative protein from
microalgae with higher yield
and lower environmental
impact.
Addressing sustainability
challenges in the nutrition
industry.
Potential to revolutionize
protein production.
AlProtein
Developing the first general-
purpose protein language
model for engineering
antibodies.
Aim to increase antibody
yield and performance
through AI-driven protein
engineering.
Implications for drug
development and
biotechnology research.
Proteinea
8.
9. Vision & Mission
Rology's mission to democratize the
radiology sector in the MEA region by
revolutionizing teleradiology systems.
Focus on improving access to radiology
services in countries with low radiologist
density like Nigeria and Kenya.
10.
11. Growth & Impact
Overview of Rology's growth from 2020 to
2022, especially during the COVID-19
pandemic.
Expansion to Kenya & Saudi Arabia by the
end of 2022.
Impact on improving access to radiology
services and reducing turnaround time for
diagnosis.
12. Using AI
Generative AI
Leveraging AI
Rology's use of state-of-the-art supervised learning
models to assist radiologists in annotating radiographic
images.
Rology leveraged Generative AI to write medical reports
and improve automatic matching between radiologists
and patients.
13.
14. Rology’s efforts in
improving automatic
matching between
radiologists and patients
has paid off, they went
from 38% of cases being
automatically matched in
2022 to a whopping 79%
in 2023.
Innovation using AI
15. Achievements
01
Decreasing the turnaround time
for annotating and reporting a
scan from 7.5 hours to 4 hours
with 99.8% medical accuracy.
Expansion in the KSA by their
acquisition of the Saudi
teleradiology provider, Arkan United.
Securing USD 500k from the Philips
Foundation & FDA clearance in 2023.
Saving a whopping 750 000 lives since the start of their
operations.
02
03
04
16.
17. Vision & Mission
AlProtein's mission to offer a premium
alternative protein that is eco-friendly,
organic, gluten-free, and vegan.
AlProtein's vision for creating a
sustainable solution to the global food
supply problem.
Distributing eco-friendly protein to food
manufacturers in various fields,
including vegan dairy, vegan meat, and
protein supplements.
18. Problem
Global meat production will be very low in comparison with the population
growth from here to 2050.
Global meat production has a devastating effect on climate with about 7.1
Gigatons of CO2 emissions yearly.
Global food supply has been disrupted following the 2022 Ukraine crisis.
20. Microalgae
AlPron
Leveraging AI
AlProtein’s
Technology
AlProtein's research on microalgae, specifically the strain
Lemna Spirulina, known for its high protein yield.
AlProtein's trademarked nutrient mix, AlPron, used to
create an optimal farming environment for their
microalgae strain.
AlProtein's IoT systems and AI-powered
biomanufacturing platform, which optimize farm
efficiency and returns through reinforcement learning
techniques.
21.
22. Achievements
01
AlProtein's patented protein
extraction technology, which
extracts an 80% nutritional yield
into their powder product.
Use of Brackish water (a mix of river
& sea water) with 85% retention,
eliminating the need for fresh
water.
AlProtein's impact on the nutrition
industry, supporting 6 UN Sustainable
Development Goals.
Securing grants totalling more than USD 150k from
UNICEF and the European Union and more.
02
03
04
23.
24. Vision & Mission
Proteinea’s vision to revolutionize
antibody engineering and accelerate
drug discovery by leveraging cutting-
edge AI technologies.
Proteinea's focus on researching
antibody structure to engineer
antibodies more efficiently and with
higher performance.
25. Challenges
An antibody protein is formed of hundreds
of amino-acids, with each of them having
the possibility to be one of 20 amino-acids.
Traditionally, pharmaceutical companies
use a trial & error process to engineer
antibody proteins.
Antibody proteins are formed out of 2
areas, the Fab area & the Fc area. Currently,
there is very low innovation in Fc
engineering. This is where Proteinea comes
in.
27. Ankh
Data
Fc Engineering
AI & Innovation
Proteinea joined the IndieBio accelerator and their
development of the Ankh protein language model.
Proteinea has collected a huge protein database, the
largest in the world, used to train their AI models.
Proteinea has developed a suite of products for Fc
Engineering, SubQer, Tuner and RaGene.
RaGene
Proteinea's AI-powered gene optimization model for
yield and quality enhancements. Results of RaGene
optimization, with up to 890% increase in yield
compared to state-of-the-art optimization tools.
28. Results of protein variants provided by RaGene vs. State-
of-the-art optimization tools.
29. But why?
Diabetes patients use insulin when their blood-sugar levels are low.
A lot of these patients take insulin in the form of injections.
Insulin is a protein that has high yield per milligram.
Due to the high yield of insulin, patients can take the injection at
home and do not have to be administered to a hospital to take the
shot.
Patients of other diseases such as Hemophilia A for example have
to be administered to a hospital to get injected with Factor VIII due
to it’s low yield per milligram
Proteinea’s efforts in increasing yield per quantity in proteins is
going to change this significantly.
This is just one example of how Proteinea’s research and innovation
is going to alter the pharmaceutical world.
30. Achievements
01
Creating the first general-
purpose protein language
model surpassing models
created by Meta AI.
Securing a $1M one-year contract
with biotechnology giant
ThermoFisher
Securing USD 2.5M from Google
Innovators in grants.
Engineering protein variants that have close to 900%
more yield than variants engineered by current state-of-
the-art optimizers.
02
03
04