Indic AI Foundation

As per World Bank, over a billion people, about 15% of the world's population, have some form of disability. Persons with disabilities are more likely to experience adverse socioeconomic outcomes than persons without disabilities, such as less education, poorer health outcomes, lower levels of employment, and higher poverty rates.

Indic AI Foundation focus is on solving challenges that prevent education and employment mainstreaming of differently abled individuals. We are proposing affordable high-quality education using online AI & Tech solutions to make it accessible and  scalable in a cost effective way. 

Our solution will help the disabled to achieve their highest education potential and will enable their inclusion and employment.

Problem: Despite 5% reservation in public sector employment, the share of employment of differently-abled in all posts remains at 0.44%, that too mainly in Grade C and D jobs. The employment share in private sector is also estimated at a poor 0.3%. Most of these seats are vacant because of unavailability of suitable and skilled candidates. 

Root Cause: The curriculum and learning materials adapted to the special learning needs of differently-abled - both in content and format - are not readily available. Mainstream content is typically not in an accessible format. Infact, most of the Learning and Education platform of training providers are not friendly to people with disabilities. Differently-abled individuals, therefore, are at a disadvantage as compared to other candidates for academic performance, competitive exams as well as employment. 

Inspiring Innovation: Our objective is to develop Tech and AI solutions to make the mainstream content accessible to differently-abled individuals and aid in skill development required for academics, exams and professional excellence.

Our main projects include developing and deploying following products:

1.Indic AI Education Platform: A disabled friendly education platform that provides accessible mainstream content. The content would be: 

a) Open: Free content is sourced from our collaborations with leading education institutes in India. The content is made accessible and then provided on our Education Platform. The content is designed for self-learning of academic, professional and job-oriented courses. 

b) Pro: Post self-learning courses, the students who require advanced training may enrol for pro courses that provide focussed training for specific competitive exams or job profiles. These are paid courses that we would subsidise using donations and CSR funds. 

2.Indic AI Access Tech Solution: We are using Deep Learning and Technology based solutions to reduce the cost footprint of making content accessible and to make our offerings scalable. For this, we are training Deep Learning models to make digital audio, video, ebooks and test series content accessible. Our models are trained on domain-specific Indian education content, ensuring our models have low error rates as compared to commercial models on the local content. We plan to use our models in-house as well as license it to third parties to financially support our operations.

In Indic AI , we are working on a solution for over 1 billion disabled people, not having access to quality education. As per World Bank, about 15% of the world's population, have some form of disability. Persons with disabilities are more likely to experience adverse socioeconomic outcomes, such as less education, lower levels of employment and higher poverty rates.

We have created an accessible platform that solves the challenges of education and employment mainstreaming of differently abled individuals. Indic AI education is a free Online platform which provides high quality education for academic , professional, and job-oriented courses for self-learning of students and support resources for teachers of disabled students.

Online accessible education will ensure higher inclusion and employment for the disabled in the mainstream world. That’s one billion people looking for accessible education for their inclusion.

Obviously, we want a solution to address them all, but our initial focus is India that has 8% of the world’s disabled population and where the employment of disabled is less than .5%.  Beyond that , we are creating strong processes and using technology solutions that will help us to scale in other counties in a cost-effective way. 

Indic AI focus is on solving challenges that prevent educational  and employment mainstreaming of disabled individuals. Accessibility, availability and affordability are the biggest constraints , for the disabled to acquire education skills, which can enhance their profile to apply for high pay jobs. Indic AI is using AI and tech solutions to convert content from leading institutes into accessible format

Our platform will enhance the employability of the disabled for high pay jobs and encourage entrepreneurship by  assigning professional mentors, enhancing computer literacy and financial support. 

Indic AI Education is the only disabled friendly learning platform in India. We want to leverage the new technology and algorithms to reach a much wider audience. Further we would license our AI and technology solutions to help corporates, education institutes, course providers etc. to convert their existing content into accessible format. 

Though there are guidelines from government on making the education accessible for people with disability, most of the existing course providers’ online platforms and content do not pass accessibility tests, as they do not consider individuals with disability as a viable target market segment. 

Implementation of the Universal Design of Learning in Indian education system is also negligible. We are forming alliances and partnership to leverage the existing content and infrastructure. We are tying up with multiple NGOs to ensure high student enrolment rate, course providers to add more courses and with professionals, mentors to guide our students and corporates to provide internships and employment. We are also putting in place a unique financing structure to ensure high course completion rate.

Our platform is fully online, responsive and accessible. This allows individuals with disabilities to get access to similar content as is accessible to other individuals, thus allowing people with disabilities to compete in academics, jobs and professional growth. 

Software:
Our LMS is customized version of OpenEdx platform. The platform is accessible and responsive and thus works on both desktop and mobile. We are using our inhouse IndicAccess solution and multiple automation scripts to convert inaccessible content into accessible content and deploying it on our LMS. Our accessibility solution IndicAccess is powered by NLP models trained using NVIDIA NeMo toolkit (https://github.com/NVIDIA/NeMo) and Mozilla DeepSpeech (https://github.com/mozilla/DeepSpeech). We are presently developing a Django based front-end for IndicAccess solution. Subsequently, we are planning to license our accessibility solution to institutions that want to automate the process for making content accessible. Automation scripts are written in Python and R.

List of Software used: Customised OpenEdx based LMS, Tensorflow and Pytorch for training/finetuning Deep Learning models, Python and R for data processing and automation scripts, MLFlow for MLOps, Gmail for emails and Slack for internal collaboration and coordination.

Hardware
We are using both inprem and oncloud servers for various tasks. Our LMS is deployed on AWS. Deep Learning model training is done using inprem GPU machine. Model inference is done on personal laptops.

Technology solution and enabling technology learning in disabled are the most important part of our solution. The technology requirements vary for different types of disability.

LMS technology: is similar to Coursera, Edx and other LMS technology software, the only difference is that the student and Course uploading interface will be accessible to the disabled students. By FY 2021, FY 2022 and FY 2023 the platform will automatically make non-accessible education content accessible for visually impaired, hearing impaired and multiple Indian languages respectively.

Support functions based on analytics : Since we are serving the disabled , we need additional support services like mentoring , sponsorship, data analytics, internships, placements & doubt clearing faculty , which will also be based on Tech solutions.  For example, for scholarships we  will create an online scholarship platform similar to  buddy4study, for placements it will be like online job portals. 

The IndicAcess technology will also vary for different disabilities, but similar solutions already exist, for example Microsoft Azure for caption and transcripts, Lets talk sign for hearing impaired, but we need higher accuracy for the India specific data. Further the Indian Sign Language has only 3000 words , so training models for Indian students will need lot of data.  

Indic AI provides  easy to use , accessible education resources for the disabled students and their teachers. This includes accessible learning resources and courses for academic, professional and job preparation. Our immediate goal is to provide accessible content and mentoring to enable the disabled students to achieve their highest potential . These immediate goals are linked to our long-term outcome of providing accessible education to  the 1 billion disabled people worldwide.  

To test some of the links we carried out a pilot with visually impaired students in India. Results show that students having access to accessible mainstream content perform equally as sighted students and there is a significant improvement in their morale and academic performance.  

Indic AI Education pilot started in Feb 2020. We have 150 visually impaired active users from our pilot launch. 67% of them are using the Indic AI Pro and 34% are enrolled for Indic AI Open. As on May 2020, we have 500 plus visually impaired students enrolled with us from Tier 1 cities in India.

FY 2021 (March 2020 to April 2021): 1250 students for Indic AI Pro and 5000 students for Indic AI Open.

FY21-FY25: Our target is to reach 1.9 million cumulative enrolments covering 0.2 million students and enable 10,000 differently abled individuals to find employment. 

We aim to achieve following short-term and long-term results by end of 2020 and 2025 

2020: 

a) Hours of accessible content: 2000 hours with 100+ courses.

b) Total students: 6250+ students 

c) Revenue: USD 20,000 

2025:

a) Hours of accessible content: 50,000 hours, with 1200+ courses

b) Total students: ~1.9 million course enrollments (1.8 million in free courses, 0.1 million in Paid) 

c) Revenue: USD 1.3 million (from commission on sales of Paid courses and licensing of our Deep Learning models).

This is going to work because of following factors: 

Our collaboration approach: At organisation level, we understand that lot of resources already exist with private institutions, corporates, NGOs and collaborating with them will enable us to make our operations commercially viable. We do not intend to reinvent something which already exists. For instance, we do not want to enter content creation space but want to invest on how to make existing content accessible.

Innovation: We believe that Tech and Deep Learning would allow us to achieve our goals impact lives of millions in a cost effective way.

Core Team: Sarita has domain expertise in Education and Training sector and Neha has domain expertise in application of Machine Learning on real-life datasets.

Enabling laws in India: as per Person With Disability Act 2016, all the public and private infrastructure and education should be completely accessible for disabled by June 2019.

Increasing awareness: in companies for inclusion and diversity. Moreover, corporate support is increasing due to Corporate Social Responsibility (CSR) Act.

We envisage following challenges in our journey:

Academic Collaboration for Content: The leading training institutes might be apprehensive to share their content with Indic AI, as they may not connect to the cause or may fear content leakage to other students etc.

Technology and AI solutions: Our biggest challenge is creating/finding datasets for developing and fine-tuning deep learning and machine learning models.

Change in orientation: The biggest challenge is to transform the thought process of differently-abled students. Due to reservations in the government sector, most of them aspire for Government jobs only and the exams conducted for them are not in accessible format. We want to empower our students to be able to find jobs in private sector and excel in their job performance.

Managing dropout rate  would be a key challenge as courses are online.

Financial: We will be dependent on external cash flows initially as our internal operating cash flows would be negative.

Schools: It will be difficult to convince schools , to allow external institutes to set up the online infrastructure for English, maths, science and logical reasoning classes. 

To mitigate the above barriers, we will use the following approach  

Academic collaborations for content: India is experiencing a big surge in the online and self-learning training portals resulting in training providers offering 10% - 50% of their content for Free. Our first target is to make this content accessible. As the content is free, training providers would be less hesitant to provide us the same content for our accessible learning platform.

Technology and AI solutions: we are using open-source datasets and data provided by our academic collaborators to train algorithms.

Change in Orientation: We would be offering large number of courses to improve job-oriented skills such as Digital Marketing, Accounting, Coding, Computer Literacy etc.

High dropout rate/Low course completion rate: We are designing and implementing a mentorship program to ensure that first-time students on our platform have access to a mentor who can personally guide them.

Financial: Though we are a NGO, we want to be self sufficient by generating inhouse revenues to support our organisation in long run. We would follow a hybrid approach by generating inhouse revenues and raising grants/CSR funds. In short and medium term we need grants and CSR funds to support our operations. By FY26 we would become cash profitable. Further, most of our expenses are variable in nature, thus we can plan our operations based on the funding raised from CSR and grants.

Schools: It is difficult and expensive for school to find good teachers, so they should find our solutions useful. 

Not applicable

We  have 3 full time and 2 part time employees.

Full time:

Sarita Chand, Co founder and CEO is working full time since Sept 2019.

Moumita, Assistant manager for Learning management system , is working since Dec 2019.

Satbir, Assistant manager for accessibility testing is working since Dec 2019.

Part Time: 

Neha Shivran, co founder and Data scientist is working part time since Sept 2019.

Contractors:

DRC Systems : for designing the LMS and Edtech platform ,  since Oct 2019.

13 Llama: for social media marketing since April 2020

We are two Co-founders and as a core team , we have the required experience and passion to deliver this project.

Sarita Chand is co-founder of Indic AI and she is heading the operations, marketing and alliances stream. Prior to Indic AI,she was the Co-founder of Edupristine, one of the leading professional training institutes in India, where she was the Head of Sales and Business Strategy. Her experience of running a successful EdTech startup for 10 years, will be instrumental in managing and scaling up our operations. 

Neha Shivran is co-founder of Indic AI and she is heading the Technology and Artificial Intelligence stream. She is responsible for conceptualizing, training and deploying deep learning models, managing technical team and coordinating with technology vendors. She has over 10 years of experience in data science, analytics and financial industry. Previously, she was Director of Risk and Analytics at Lenddo EFL India, an alternate data startup. She has also worked as Head of Custom Modeling with India’s largest credit bureau. She is an Engineer from NSIT, Delhi and MBA from Cranfield University, UK.

We, therefore, have expertise in setting up, managing and scaling up operations in the Edtech startup. 

1. Training Institutes: Few of the leading training institutes in India have already partnered with us: 

a) CL Educate Limited: for Law and MBA exams. www.careerlauncher.com 

b) Vedantu: for maths and science courses. https://www.vedantu.com/ 

c) Adda247: for government exam courses. https://www.adda247.com/ 

d) Allen Career Institute: for IIT JEE preparation, and material for Maths and Science courses (from 9 to 12 class) 

e) Edupristine: for finance and accounting courses. https://www.edupristine.com/ 

f) Digital Vidya: for digital marketing courses www.digitalvidya.com 

g) Kredent Academy: for finance planning courses. www.kredentacademy.com 

h) Chandoo.org for MS Excel courses. http://chandoo.org/ 

i) Career Guide: for Career Counselling. www.careerguide.com 

2. NGOs: we have tied up with NGOs for students onboarding and for training content 

a) Vision AID for python, computer literacy and accessibility trainings https://www.visionaid.org/ 

b) Saksham NGO for student onboarding and Daisy Consortium Online courses http://saksham.org/ 

c) National Association of Blind for student onboarding. http://www.nabdelhi.in/ 

d) Blind 2 Visionary for Maths and Science Trainings http://blind2visionary.org/

We do not design special courses for differently-abled students. Rather, we convert the mainstream courses provided by leading training providers into  accessible format using AI and tech solutions. This helps our disabled students them to compete with others and achieve their full academic potential. Indic AI Education Platform provides students access to three category of courses: 

a) Academic courses: Courses for classes I to XII. 

b) Competitive exams: In India, there are competitive exams for higher education seats (such as JEE, GRE, MBA, CLAT, CA), public sector jobs (such as exams for Bank Probationary Officer, UPSC, Teaching) and private sector jobs (such as quantitative abilities, english communication, computer literacy). We are forming alliances and partnerships with training providers for competitive exams to make their content accessible to differently-abled students who are pursuing higher education or who are looking for employment.

c) Professional courses: These courses are meant for people with disabilities that are looking for skill development to either enter the job market or perform better on existing jobs. These courses cover areas such as coding, digital marketing, accounting, financial analysis etc. 

Therefore, our solution caters to development of academic as well as job-oriented skills of differently-abled individuals in age group 5 to 29 years.

The solutions are provided in online format, the students can learn using mobile phones or laptops. The Open courses are free courses , whereas marginal fees is charged for advanced training in Pro courses.

What is your path to financial sustainability?

Expert advice: we are an Edtech NGO, however we need to scale and grow like another Tech startup , so that we can  reach millions and solve the problem of education of disabled across various countries. The strategic advice from experts on board, will help us to build strong foundation in terms of processes and technology. 

Exposure to more NGOs: Solve accepts applications from social enterprises around the world at various growth stages and  learnings from fellow NGOs will help us to accelerate our growth. 

Technical support: The MIT solve team has exposure to many Tech startups, their tech expertise can help us to enhance our Tech solutions. 

Marketing exposure : Winning this award will help us financially and enhance our reputation with training partners & corporates.

Financial Support: The collaboration of MIT solve with various foundation for Prize and funding opportunities can help us accelerate our growth. 

In Indic AI, we understand that lot of resources already exist with private institutions, corporates, NGOs and collaborating with them will enable us to make our operations commercially viable. 

Academic Collaborations : with training institutes and NGOs , which can provide structured content which can be made accessible for the disabled.

Technology partners: There are many solutions for virtual education, however none of them are accessible for the disabled. We want to learn from the MIT solver startups which are providing tech solutions for digital education in schools and colleges. 

STEM: subjects are the most important to enhance the creativity and entrepreneurship in disabled students. We are looking for partners which have existing solutions for STEM education.

Distribution Network in other countries: Indic AI team is familiar of scaling business in South Asia, however we need affiliates which can help us to scale our operations in other countries. 

For STEM education : we want to work with Khan academy, MIT scratch , Code.org, Code club, Practical education network solver or similar STEM providers to provide accessible education to the disabled. 

Technology Partners: We want to learn and partner with solver Education modified, Livox, Early bird and leading edtech startups like Duolingo, Edx , Torsh, Blackboard to understand and implement new technologies in the education industry. 

Academic Collaborations: We have partnered with leading institutes in India namely Career Launcher, Vedantu, Career Guide, Superprofs, Adda247, Edupristine, Kredent, Finshiksha etc for competitive exams and professional trainings in accessible format for the disabled. By year 2025, we aim to partner with 1300 institutes to make 60,000 hrs of accessible content. 

Distribution Network: partnering with government organisations and International organisation like ILO, UN, DREDF, International disability alliance . 

UNDP (United Nations Development Programme) statistics show that in developing countries, 90% of children with disabilities do not attend school.  The picture for disabled girls is even bleaker. While the literacy rate for adults with disabilities is 3%, just 1% of women with disabilities are literate.

Though our focus is on education and employment of entire disabled population, we assign higher priority to our female students, while assigning mentors, STEM courses,computer literacy, entrepreneurship trainings and financial support. 

As an organisation, both our founders are women entrepreneurs and our company female employment ratio is 75% . 

The  innovation for Women prize will help us in the development, implementation and scaling up of following: 

1. Developing and deployment of our Deep Learning (DL) Models to automatically make mainstream content accessible for the disabled. 

2. Invest in our Education Platform: We would invest in our Education Platform to take it from Beta to stable release version, with focus on testing, bug-fixing and incorporate feedback received during pilot test. 

3. Invest in business operations: We need to scale up our operations to reach our target of 60000+ hours of content and 2 million users by 2025. Specifically, we would use funds to hire additional team members for following tasks: a) Alliances team to tie-up with NGOs (to onboard students) and training providers (to source content) b) Machine Learning team to train our Deep Learning models c) Accessibility testing team to do Quality Assurance for the content uploaded on our platform

As per United Nations, in developing countries, 80% to 90% of persons with disabilities of working age are unemployed, whereas in industrialized countries the figure is between 50% and 70%. 

In Indic AI, our vision is to empower disabled people around the world with high quality education for their inclusion and employment in high pay jobs. We have developed fully online, tech based , disabled friendly education platform,  specifically focusing on academic , competitive exams and job-related professional courses . 

Further we encourage entrepreneurship and provide access to funding, mentorship and professional trainings. 

The  GM prize will help us in the development, implementation and scaling up of following: 

1. Developing and deployment of our Deep Learning (DL) Models to automatically make mainstream content accessible for the disabled. 

2. Invest in our Education Platform: We would invest in our Education Platform to take it from Beta to stable release version, with focus on testing, bug-fixing and incorporate feedback received during pilot test. 

3. Invest in business operations: We need to scale up our operations to reach our target of 60000+ hours of content and 2 million users by 2025. Specifically, we would use funds to hire additional team members for following tasks: a) Alliances team to tie-up with NGOs (to onboard students) and training providers (to source content) b) Machine Learning team to train our Deep Learning models c) Accessibility testing team to do Quality Assurance for the content uploaded on our platform

In Indic AI , we are working on a solution for over 1 billion disabled people, not having access to quality education. Indic AI education is a free Online platform which provides high quality education for academic , professional, and job-oriented courses for self-learning of students and support resources for teachers of disabled students. 

We further enhance the digital literacy of the disabled which makes them independent and enables  leapfrog growth in their learning experience . For example our digital literacy trainings for visually impaired students, enables them to have access to screen reader compatible learning resources on our platform.

The  Gulbenkian award will help us in the development, implementation and scaling up of following: 

1. Developing and deployment of our Deep Learning (DL) Models to automatically make mainstream content accessible for the disabled. 

2. Invest in our Education Platform: We would invest in our Education Platform to take it from Beta to stable release version, with focus on testing, bug-fixing and incorporate feedback received during pilot test. 

3. Invest in business operations: We need to scale up our operations to reach our target of 60000+ hours of content and 2 million users by 2025. Specifically, we would use funds to hire additional team members for following tasks: a) Alliances team to tie-up with NGOs (to onboard students) and training providers (to source content) b) Machine Learning team to train our Deep Learning models c) Accessibility testing team to do Quality Assurance for the content uploaded on our platform

As per World bank, there are 1 billion disabled people in the world, who experience adverse socioeconomic outcomes, such as less education, lower levels of employment and higher poverty rates.

Our vision is to empower disabled people around the world with high quality education for their inclusion and employment. We have developed fully online, disabled friendly education platform,  specifically focusing on academic , competitive exams and job-related professional courses .

Our solution is fully online, tech based and fully accessible, which differentiates us from NGOs and other training providers. At organisation level, we understand that lot of resources already exist with private institutions, corporates, NGOs and collaborating with them will enable us to make our operations commercially viable. We do not intend to reinvent something which already exists. 

We have trained  deep Learning Models  on educational content to automate the process of making content accessible, which helps us to scale our solution in a cost effective way. Our solutions can be used directly by the students for self learning or by teachers to enable their students.

Further we would license our AI and technology solutions to help corporates and education institutes,  to generate revenues to be sustainable organisation. 

Even during Covid, many countries are encouraging digital literacy platforms, but none of these platforms are accessible for disabled. 

The  Patrick J. McGovern AI Humanity prize will help us in the development, implementation and scaling up of following: 

1. Developing and deployment of our Deep Learning (DL) Models to automatically make pre-recorded video/audio, text , questions content accessible. Training and deploying DL models requires a) Access to data b) Model Training c) Model Deployment. The award will help us in tying up with a large number of training institutes in and give us access to their learning material. This would augment the base data for our model and reduce its error rate. The challenge will also help us invest in IT Infrastructure (specifically GPUs) to train and deploy our models at scale. 

2. Invest in our Education Platform: We would invest in our Education Platform to take it from Beta to stable release version, with focus on testing, bug-fixing and incorporate feedback received during pilot test. 

3. Invest in business operations: We need to scale up our operations to reach our target of 60000+ hours of content and 2 million users by 2025. Specifically, we would use funds to hire additional team members for following tasks: a) Alliances team to tie-up with NGOs (to onboard students) and training providers (to source content) b) Machine Learning team to train our Deep Learning models c) Accessibility testing team to do Quality Assurance for the content uploaded on our platform