InclusionHub, a subsidiary of SeqHub

Machine Learning (ML) is powering the digital economy; however, the lack of diversity in the authorship and stewardship of ML models leads to bias in its algorithms. Groups underrepresented in the field-- women, Black, Latinx, and LGBTQ+ individuals-- are particularly vulnerable. When these algorithms are deployed in high-stakes decisions in society or healthcare, racial and gender discrimination is inevitably codified to perpetuate these inequities. 

 To make diverse candidates eligible to influence algorithms and connect them to employers, our solution, InclusionHub, has three goals: 1) to increase the pipeline of coders from underrepresented groups in high school, 2) to provide data science training in advanced ML, and work experience to make them attractive to employers and 3) a digital leadership program to train ML practitioners in business and entrepreneurship so they can solve the problems that impact their communities resulting from model bias and other pernicious data practices.

The specific problem we are addressing is the codification and perpetuation of discrimination in the digital economy resulting from a lack of diversity in the data science workforce[1]. This global issue affects all individuals, especially those who identify as women, Black, Latinx, and LGBTQ+, typically underrepresented in the digital economy[2]. The dearth of representation and participation among these groups is a consequence of lack of access to necessary training and apprenticeships required to excel in data science fields[3], and related structural barriers: limited data literacy curriculum offered at public schools, challenges identifying post-graduate jobs, difficulty gaining access to the skills required for promotion once employed. 

The tech industry adds ~9,600 jobs to the economy every month in the US [4]. Recently the OECD warned that long-term unemployment could remain high in wealthier countries because of a mismatch between open positions and workers’ skills[5]. There is a skills shortage of data analysts[6] and machine learning practitioners[7] , and employers express the need for “work-ready” data scientists. Training more practitioners, especially those underrepresented, will help meet employer needs and mitigate algorithmic bias.

And finally, empowering local entrepreneurs with tools to solve their local problems themselves[8].

Our solution is 3 nested programs targeted at those currently underrepresented in data science and most affected by algorithmic bias.  Specifically, we train underrepresented populations in data science, find them employment at their highest-performing level, and make them specialists in ML so they can influence the direction of the digital economy. 

We start with the opportunity gap experienced by high school students with our College Ready program. By increasing access to coding, we can fuel the pipeline of future leaders and ML practitioners.  

Second, our Work Ready program encompasses data science foundations focussed on ML, business modules like design thinking, critical thinking, and storytelling with data; it emphasizes ethics and social implications of ML. On completion of training, all students will work on real-world problems in a safe learning environment - an apprenticeship model under the supervision of experienced lead data scientists. Our assessment tools will measure when they are “work-ready.” We will identify employment for participants via our network of business sponsors.  

Our third program provides data science leadership that will accept our own graduates and other candidates to prepare them to be thought leaders in the business where they can influence the future direction of the digital economy.

Bias in Machine Learning (ML) algorithms potentially impacts everyone but has been shown to have a more egregious impact on marginalized populations, particularly in high-stakes situations - sentencing decisions, facial recognition, law enforcement, access to credit, insurance rates, hiring processes and employment opportunities[1]... and in disparate health care and outcomes. For example, in personalized medicine where algorithms are core to genomics research, most diseases researched to date are focused on populations of European descent. The underrepresentation of scientists from diverse ethnicities and environmental backgrounds may lead to the absence of hypotheses and priorities that experts with different perspectives would bring to genomic research.[2]

We want to increase diversity in ML practitioners by focusing on women, Black, Latinx, and LGBTQ+. This specialized field requires a foundation in, first algorithmic coding, and then data science with ML specialization.

From our customer discovery interviews, we identified that our programs will benefit high school students in public schools by offering digital skills they need for the new economy. Undergraduates trying to find internships that enhance their resumes to compete for future career opportunities. Graduates who did not attend an employer-preferred college struggle to find employment without work experience. Underrepresented populations in data science who are underemployed in their current roles and feel they are passed over for promotion have a desire to learn leadership skills.

Machine Learning is powering the growth of the digital economy and requires more diversity. During our customer discovery interviews, we identified that companies and research institutions are experiencing a shortage of qualified practitioners and simultaneously a need for diverse candidates as part of their competitive advantage. Millennials and GenZ  (recent entrants into the job market) are increasingly looking for employers with demonstrable DEI policies[3], and companies are recognizing that diversification into new products and services requires a diversified workforce.

Bias in ML algorithms disproportionately affects the underserved and underrepresented population in the stewardship of these algorithms. Our target population is those currently underrepresented whom we want to train and provide work experience to increase the diversity of algorithmic coders, data scientists, and, ultimately, ML practitioners. We will obtain skills requirements in consultation with employers, and our assessment platform will ratify work readiness to remove barriers to employment. Our leadership program will continue to train them to advance to their highest potential. Diversity in leadership has been shown to impact the retention of minorities in data science positively.

SeqHub is a boutique data science consulting company that drives insights for companies through interdisciplinary research and applications of machine learning and AI.

InclusionHub is being incubated in SeqHub Analytics, where we have been testing the curriculum and the apprenticeship model in the US and Nigeria - 10 high school students in the US have trained, and 6 graduate students in Nigeria. 

Through customer discovery, we have identified businesses and research institutions willing to sponsor IncusionHub by paying for us to perform projects under the apprenticeship model, employing our graduates, and using our advanced training modules to enhance their DEI efforts.

Outreach to educational institutions and employers has confirmed the need for our apprenticeship model for work experience to help underrepresented graduates secure their first job and improve their promotion opportunities with our advanced training. 

As an immigrant and a Black data scientist and educator, I am well-positioned and qualified for the ServiceNow US Racial Equity Prize. Nurturing and developing young talent is core to my purpose and is something I have been privileged to learn to do quite effectively. As the architect of a thriving Computer Science program at Pierrepont School, a K-12, I can make complex topics accessible at lower levels. As an academic and research advisor to college undergraduate and graduate students, I am experienced in helping students identify and develop technical and non-technical skills they need for their different academic and research interests. As a data science consultant to companies, I have experience identifying and targeting real business value. Finally and most importantly, I have surrounded myself with years of complimentary experiences in my co-founder and team as a whole. Altogether, I am confident that I am well-positioned to lead our mission at InclusionHub to train underrepresented populations in data science, find them employment at their highest-performing level, and make them specialists in ML and AI so they can influence the direction of the digital economy.

100% of our funds will support our mission, and I would particularly welcome the opportunity to focus my diverse team on developing a robust set of assessment tools. These tools are pivotal to advancing equity and cultivating entrepreneurship.

Our operating model innovation combines data science training, targeted business skills in a safe, simulated work experience - an apprenticeship model to solve actual business problems submitted by our sponsors. Our sponsors will sign up to perform “the client” role and provide feedback as part of the simulation. This high-impact simulation will fill the gap of work experience that employers are demanding for recent graduates. Simply diversifying the data science workforce is insufficient to promote equity and inclusion; we will require our sponsor companies to demonstrate their commitment to promoting equity and inclusion in their organizations. In its final state, our assessment tool will measure attributes of inclusion like the psychological safety of teams.

The network of College Ready, Work Ready, and Leadership Ready programs combined address the overall underrepresentation in machine learning and AI. In addition, our partner Tiny & Great provides a feeder system of grade 8 coders, laying a solid foundation for those underserved to achieve the necessary proficiency to participate in the College Ready Program. 

Our integrated diagnostic and performance assessments are also innovative. Rather than only diagnose the gaps in our students, our assessments target the gaps in our program and use a combination of sponsor project descriptions, sponsors’ feedback, students’ feedback, and students’ performance for continuous refinement. 

In the final state, InclusionHub will partner with sponsors and investors to open incubators to support our entrepreneurial graduates in developing solutions for local challenges.

The key value of our approach comes from combining a broad and highly practical curriculum with an Assessment Dashboard to provide a targeted training experience to our students and monitor the impact of our training over time. We are currently able to assess our students' technical skills - algorithmic coding, data science, and machine learning - to ensure they enter at the correct level of the curriculum and exit with the necessary technical skills.

We are actively working on expanding to target other important areas. For example, in partnership with Blade Kotelly, we plan to develop how to assess: invention and leadership skills, self-efficacy, concept generation, and other innovation dimensions. 

In partnership with academic and industry experts, we collaboratively design assessments for other required skills like business analytics abilities, data science product management, understanding of psychological perspectives behind decision making, AI literacy, and ethics to be implemented and integrated into a unified diagnostic dashboard.

This unified dashboard of assessments will allow us to easily see the areas of strengths and weaknesses of each student across several dimensions.

Over time, assessment results will become a source of data to effectively scale our ability to map tiers of competency with appropriate data science roles, further refine the design and implementation of our programs and their assessments, and monitor and promote a psychologically safe experience for students. 

While initially targeted to our focus in data science, this assessment dashboard provides easy adaptation for other technical/STEM fields that sorely lack diversity.

The technical assessments we currently use are commonly deployed in the technology industry. 

We are leveraging our partners' (Blade Kotelly - BK) thought leadership for our innovation and leadership assessments and adapting their market offering. 

BK’s quantified innovation assessment is a proven method to quantify and boost the innovation skills and culture in organizations. Our partnership with this data-driven innovation company provides us with validated industrial metrics to measure and track our apprentices' growth and the efficacy of our program. The BK quantified innovation methodology is currently used by companies like Hasbro, VMware, to mention a few.

Our innovation combines these proven assessment methods above. It will expand on it to create a unified Assessment  Dashboard that measures and tracks the other dimensions of data scientist work-readiness and leadership.

To achieve this next generation, The IncusionHub Assessment Dashboard, we will combine proven assessment tests and techniques used by our assembled team of experienced Social, Organizational & Behavioral Psychologists, Educators, Business leaders, practicing data scientists, and expert ML practitioners.

Data privacy and security concerns that may arise as potential issues of an assessment dashboard will be mitigated by our compliance with the General Data Protection Regulation (GDPR). We demonstrate our commitment to the highest level of privacy and security, starting with a privacy policy that details and explains how our organization uses the personal data we collect from all participants in our program. The purpose of the InclusionHub Dashboard will be for InclusionHub to monitor itself, measure our impact, and refine our offerings. We do not plan to be sharing individual data with anyone outside our organization.

Another potential risk raised by our approach to assessments is the possibility of missed precision. If the assessments are not aligned appropriately with measuring valuable areas/outcomes, they may focus on the wrong areas. To mitigate this risk, we are using a data-driven approach, focusing on first principles. We start by identifying the key areas that we believe are important for students based on our experiences, choosing/developing assessments to measure those areas, evaluating whether our assessments are accurate and useful by working closely with our students and hiring organizations, and finally re-assessing the foundational areas of focus by measuring and observing the successes and challenges of our students after they complete our training programs.

Our pilot College Ready program in 2021 trained ten high school students.

Our pilot Work Ready program trained six college graduates and undergraduates.

The target for our 2022 College Ready summer immersion program is 20.

Our 2022 goal for the Work Ready Ready Program is to apprentice 30 in the US. We have identified two potential employers who have given verbal commitment to sponsor our apprenticeships. 

In 5 years, we plan to successfully train 300 apprentices, launch our Leadership Ready Program, and expand InclusionHubs globally.

Our aspirational goal is to mitigate Machine Learning algorithmic bias and its consequences on everyone, especially those underrepresented in data science and the digital economy, who are most adversely impacted. Increasing representation in data science, specifically, the specialization of machine learning at the entry-level should have a knock-on effect on authorship and stewardship of these ML algorithms and AI models. It will also increase recruiting and retention in future generations.

We intend to support Incubators in local communities that are best-positioned to provide digital innovations and solutions to their local problems.

We will explore expansion to further InclusionHubs leveraging our training materials, operating model, and technology.

• Number of high school graduates who attend the college of their choice

• Number of graduates who are “ready to work”

• All our graduates are from an underserved community

• Number of committed paying clients (sponsors)

• Number of apprenticeship projects

• Income from Sponsors

• Number of Incubators

• Number of global InclusionHubs 

InclusionHub is a subsidiary of SeqHub Analytics committed to operating at zero profit.

We currently have 1 full-time team member, 10 part-time team members, and 4 advisors.

Our team is a mix of educators, data scientists, machine learning practitioners, proven business leaders, psychologists, and organizational change experts focusing on leadership and leadership training. This enables us to match the curriculum with the employer's needs. 

The proven business leaders and organizational change specialists understand the success criteria of new entrants to the workplace and the skills required to navigate the environment to become successful leaders. Likewise, the data scientists and machine learning practitioners understand the componentry of models to build for less bias and adjudicate fairness in ML algorithms.

All are seasoned educators in either high school, college, or work-based learning. All are passionate about diversity, equity, and inclusion and have the first-hand experience on how access to training and work experience can impact a career.

 The team also has specialists in leadership assessment and technical skills assessment.

The model profile for our leadership team is embodied in Taiwo Togun. Taiwo is an experienced and passionate educator and data scientist - a high school teacher who has led successful academic and apprenticeship programs targeted at high school and college students. And at SeqHub, he leads an interdisciplinary data science consulting team that drives insights for companies through the application of interdisciplinary research, machine learning, and AI. He holds an Executive Masters in Technology Leadership from Brown University and a PhD. in Computational Biology & Bioinformatics from Yale University. 

Diversity is the driving force behind our team and at the heart of our commitment to success. Our leadership team’s own diversity reflects the diversity of the population we aspire to serve. Selected for their competent technical skills, ability to draw from their own experiences, and united by the desire to contribute to the necessary changes in the workforce that will be inclusive. We share the belief that talent is to be found in underserved communities, and all that the talent requires is training, mentoring, and opportunities.

Psychological safety and self-management are core to our HR practices, which encourage inclusivity in contribution and decision-making.

Goals

• 80% of our leadership will always be an under-represented minority; black, Latinx, women or LGBTQ+

• 100% of our apprentices will always come from underserved or underrepresented communities

We believe combatting algorithmic bias and its effects is everyone’s responsibility and particularly experienced machine learning practitioners. SeqHub currently collaborates with the Boykin Lab at Brown, where they are evaluating fairness perceptions that influence ML algorithms. 

When we heard about the purpose of this Challenge is to “leverage technology and strengthen digital capacity-building to meet the needs of the current and future digital workforce” we recognized that we share this mission with the added focus of increasing diversity in machine learning to promote fairness in algorithms. It motivated SeqHub to leverage its in-house talent and to formally launch InclusionHub.

Our initial purpose was to get the word out there, to highlight our viable solution to mitigate bias in algorithms. We welcome the networking opportunities this program offers to scale our training to those without access to this important growing sphere of the digital economy.

Our activities include:

- a Work-Ready program pilot for 6 College undergraduates and graduates.

- successful summer College Ready program pilot

- partnership with Tiny & Great and Pierrepont School to teach coding to students from the underrepresented population at public schools.

-active research collaboration on the perception of fairness in ML algorithms

PR - We want to shout out that a powerful way to combat bias and discrimination in frontiers of the digital economy is to have diversity in the authorship and stewardship of machine learning algorithms. To level the playing field of healthcare and advances in precision medicine, one needs diversity in those directing research areas and making the funding decisions.

Assistance in networking with corporations to sponsor underrepresented candidates to further their DEI agendas - both for entry-level hiring and training new and current employees in machine learning and leadership.

Outreach to organizations (educational, community, social, etc) to help us identify candidates for ML  training and AI model development.

In the future, as we scale, connections to recruiting agencies that share our DEI mission to assist with the placement of qualified candidates.

To network with potential partners who may be named above as our competitors. For example, P-TECH, whose program offers data science at our target schools simultaneously with an associate's degree. Our machine learning specialization could be an excellent augmentation to their offering, and they could provide their graduating students the opportunity to further specialize in NLP and conversational AI.

Local business cooperatives that share our mission and goal. For example, Stamford Partnership has a tech program and offers co-funding for initiatives that transform cities, especially using technology.

Educational partners that will assist us in developing curricula and assessments, such as Pierrepont Schools. Industry partners that will assist in similarly developing educational and assessment materials to ensure that graduates are positioned to have a significant social impact in the fields of ML and AI.