Month: November 2024

In the last quarter of 2024, the Hubtel AI lab collaborated with Hubtel’s management team to develop “HubtelWay,” a bespoke artificial intelligence platform designed to deliver up-to-date information. HubtelWay is shorthand for Hubtel’s unique organizational culture, policies, and leadership practices. This system is a chatbot aimed at assisting over 700 Hubtel employees. It answers specific questions and addresses their concerns without the need for any human involvement.

The use of in-house chatbots within organizations is steadily gaining popularity. For example, Deloitte recently announced the deployment of an AI chatbot to support employees in audit and assurance tasks. JP Morgan and Morgan Stanley have developed chatbots to assist with research analysis and high-end customer financial advice. These advancements reflect a shift in the conversation around artificial intelligence—from building products for external consumers to finding high-value utility within the company for employees.

Esther Crawford, Director of Product at Meta, recently shared this on Threads to celebrate the launch of Metamate, an internal AI chatbot for Meta’s employees: “You’ve never used Metamate if you don’t work at Meta. It’s an AI for employees that’s trained on an enormous corpus of internal company docs. I use it all the time for efficiency gains. Any sizable company operating without an internal AI tool is already behind the curve.”

At Hubtel, we agree! Building custom chatbots for internal knowledge sharing not only supports transparency around policies and regulations but is also a smart way to cut down on Human Resource management costs. Especially for fast-growing companies like Hubtel, having a scalable system that can support an expanding team is crucial.

How HubtelWay AI works

HubtelWay AI operates using Retrieval-Augmented Generation (RAG) with OpenAI models to process our internal knowledge base (the ‘corpus’ of documents). The system breaks down company documents into smaller, focused chunks of text. Using an embedding model, these chunks are converted into unique vector embeddings—essentially numerical representations that capture the meaning of each piece of content. These embeddings are then stored in a vector database for efficient retrieval.

When an employee asks a question, HubtelWay AI converts the query into the same vector format. It then finds the most relevant document chunks. These chunks provide context for the Large Language Model (LLM) to generate responses based strictly on Hubtel’s documentation.

Data integrity and privacy are essential in such a system. The system is designed to follow Hubtel’s data policy, which maintains data confidentiality and protects employees’ privacy. Personal data or sensitive employee information is not stored or shared, as the chatbot only accesses anonymized and non-personally identifiable organizational content to provide responses.

Practical Applications and Future Enhancements

Imagine a new team member joining Hubtel. Instead of spending weeks navigating various documents, they can simply ask HubtelWay AI questions like, “What are Hubtel’s core values?” or “How does Hubtel approach teamwork?” This immediate access to cultural knowledge allows them to align with our ways of working from day one.

For current team members, HubtelWay AI offers reliable support for day-to-day decisions. Whether refreshing their understanding of company policies or checking details about our leadership structure, the system provides consistent information across all our locations. This uniformity is essential as we expand across cities.

As we look towards the future, HubtelWay AI may incorporate more personalized features. These may include department-specific knowledge or project-based frequently asked questions (FAQs) to better cater to specialized teams. In addition, we are exploring the integration of real-time updates. This feature would allow the chatbot to reflect any new policy changes instantly. This way, employees will always have access to the most current information.

Embrace the HubtelWay…

We encourage all ‘Hubtelites’ to make HubtelWay AI their first stop for questions about the company. Whether you’re a new team member finding your feet or a seasoned ‘Hubtelite’ exploring new aspects of our organization, HubtelWay AI is your gateway to a deeper understanding of the Hubtel Way—safeguarded by technology that values both efficiency and privacy.

Reaching the stage where tackling scaling challenges is part of your daily hurdles is a great place to be. It’s in one way or the other, the kind of problem every innovation team aspires to face. It’s a sign your innovations are succeeding, but it also requires evolving your teams and systems to meet new challenges and opportunities. Scaling an engineering team from a modest 16 to over 100 engineers is a significant milestone for any company, and at Hubtel, this journey has been one of excitement, learning, and growth. It reflects our commitment to building not just an engineering powerhouse, but a culture and ecosystem where innovation thrives.  

As a fintech company heavily engaged in e-commerce and building systems that significantly impact both our operations and many key businesses in the market, our ability to scale quickly became a critical component of our success. 

Let’s explore the strategies, challenges, and lessons learned in scaling Hubtel’s engineering team to meet the demands of an evolving industry, complex projects, and the need for a robust technological infrastructure to support our growing suite of products and services. 

1. The Need for Scaling: Recognizing the Inflection Point 

Scaling Hubtel’s engineering team was not a decision made overnight. It was the result of careful consideration and an assessment of where the company was headed. In our early stages, the engineering team was a lean 16-person group that handled all aspects of product development, from coding and infrastructure to bug fixes and deployments. This worked well when our operations were focused on relatively limited product offerings. 

However, as we expanded our vision to “Drive Africa forward by enabling everyone find and pay for everyday essentials” and widen our product scope, particularly into more complex B2B services, it became clear that we needed more hands-on-deck. We were reaching an inflection point where our current capacity couldn’t keep up with the new projects and demands. Our ambitions to deliver a best-in-class fintech platform required us to reimagine what our engineering team would look like in order to meet the technical and operational challenges ahead. 

2. Laying the Foundation: A Clear Framework for Growth 

Before jumping into the hiring frenzy, we needed to put a framework in place. Scaling is not just about adding more people, it’s about hiring the right people, at the right time, for the right reasons. The following were our foundational pillars for this process: 

Defining Roles and Responsibilities 

The first step was defining what roles we needed to fill. Let me say that we have long mastered the practice of maintaining a clear separation of concerns in Hubtel engineering frontend, backend, mobile, Infrastructure, with each team operating with specific objectives. But in a smaller team, engineers often wear many hats, juggling between multiple projects and responsibilities and while this flexibility is valuable in a small setup, we recognized that to scale effectively, we had to specialize. 

We clearly defined product categories and formed a full engineering team around each product (frontend developers, backend engineers, machine learning engineers, data engineers, DevOps specialists, quality assurance (QA) engineers, and product managers). This specialization allowed each team member to sharply focus all their expertise on a single product and give their best, giving them a singularity of focus, improving both the speed and quality of work. 

Establishing Core Competencies 

We identified core competency areas that every engineer needed to excel in. These included both technical skills, such as proficiency in key programming languages and frameworks (e.g., C#,.NET, JavaScript, Reactjs, Vuejs), and soft skills like communication, problem-solving, and collaboration. Since engineers often work across teams, strong cross-functional communication became essential as we scaled.  

To ensure fair, data-driven hiring decisions, we created a detailed competency matrix used in all interviews. This matrix allowed us to score prospective engineers in each competency area, ensuring we evaluated both technical expertise and cultural fit as accurately and consistently as possible. 

Creating a Scalable Hiring Process 

Hiring quickly doesn’t mean hiring haphazardly. We knew we needed a robust hiring framework that would allow us to evaluate candidates quickly and fairly. Our hiring process was designed to be rigorous, with multiple stages of interviews: an initial talent phone screening – to show appreciation for the time the talent invested in the application and explain the journey ahead, an online assessment that tests both technical, cognitive and soft skills, and a physical interview followed by a coding challenge that reflects real-world problems engineers would face at Hubtel. 

3. Attracting Talent: Building a Pipeline of Engineers 

One of the biggest challenges we faced in scaling Hubtel’s engineering team was finding the right talent. As competition for skilled engineers has grown globally, especially in the fintech space, we needed to think creatively about how to attract talent. 

Branding Hubtel as a Destination for Engineers 

Building a strong employer brand became crucial during this phase. We focused on positioning Hubtel as a place where engineers could tackle complex, meaningful challenges. This was largely achieved through intentional leadership, fostering a positive internal culture, and benefiting from word-of-mouth referrals, as our engineers and network shared their positive experiences, attracting top talent organically. They highlighted the exciting work we were doing in fintech and how they were driving impactful changes in revenue mobilisation  in the utility sector, instant services, e-commerce etc. 

Partnerships with Universities and Coding Academies 

Another key strategy for talent sourcing was building strong relationships with educational institutions. We partnered with many universities and coding academies, such as the University of Ghana (UG), University of Mines and Technology (UMaT), Ashesi University, and Bluecrest College, among others, to create a pipeline of young, enthusiastic engineers. Through internships, workshops, and mentorship programs, we not only gave back to the community but also developed a steady stream of new talents that could be integrated into our team. 

Referrals and Word of Mouth 

Word of mouth became one of the most powerful tools in attracting talent. We incentivized our existing employees to refer talented engineers to us, creating a culture where our engineers felt proud of working at Hubtel and excited to bring others into the fold. 

4. Talent Development: Onboarding and Growing Engineers from Within 

Hiring engineers is only part of the equation. Developing talent from within has been equally important in our scaling journey. One of our core beliefs at Hubtel is that a great engineer is not just one who knows how to code, but one who continues to learn and grow, especially in leadership. 

Onboarding Mentorship and Knowledge Sharing 

Our onboarding process is a crucial first step in integrating new hires. We’ve designed a comprehensive structure to ensure that new team members quickly familiarize themselves with our tools, processes, and way of working in an immersive and efficient manner. Following this, new hires are paired with senior engineers who provide ongoing mentorship and support, helping them deepen their understanding and adjust smoothly while reinforcing a culture of collaboration and teamwork. 

We also created a knowledge-sharing culture, where engineers regularly share what they were working on, new technologies they were exploring, or challenges they were facing. This kind of open communication not only increased the overall knowledge base of the team but also built a sense of camaraderie. 

Continuous Learning and Development 

To achieve our goals, we invested heavily in learning and development programs ranging from paying for online courses and certifications to organizing an in-house training academy (Hubtel Academy). By providing engineers with opportunities to enhance their skills and explore new technologies, we ensured they remained at the forefront of the industry. At the same time, this approach helped them build the capacity to integrate and lead new team members, fostering a smooth transition into our culture. 

5. Building a Culture of Collaboration and Innovation 

One of the challenges of scaling is maintaining the culture that made you successful in the first place. . During the pivotal phase of scaling Hubtel’s engineering team, we were intentional about preserving our culture of learning, collaboration and innovation. This wasn’t just about ensuring that everyone worked well together; it was about fostering an environment where people were excited to bring new ideas to the table without promoting speculative import of culture. 

Product-Specific Teams  

One of our greatest innovations was transitioning from cross-product teams to product-specific teams. This played a key role in scaling Hubtel’s engineering team effectively. By having engineers, product managers, designers, and other professionals dedicated to individual products, we created a clear focus for each team, allowing engineers to specialize and deeply understand the product they were working on. This structure not only improved efficiency and innovation but also provided a clearer path for scaling the engineering team, as we could hire and grow talent based on the specific needs of each product. This targeted approach helped us expand the team strategically both vertically for a specific product and horizontally across the organization. 

Fostering Innovation to Support Team Scaling 

By setting up a dedicated innovation and tooling team, we created a structured pathway for engineers to contribute new ideas without disrupting their primary responsibilities. This approach not only encouraged continuous improvement but also allowed scaling Hubtel’s engineering team more effectively. Engineers could focus on their core tasks while still having a platform to propose innovations, which were then reviewed and refined by the dedicated team. As successful innovations were implemented across product teams, it improved processes and workflows, allowing us to onboard and integrate new engineers more efficiently. This structure ensured that as our numbers grew, the quality of work and pace of innovation remained high across the organization. 

6. Scaling Engineering Leadership 

As our engineering team grew, so did the need for more structured leadership. Leading 16 engineers is very different from leading 100. We needed to ensure that communication, performance, and team dynamics remained effective, even as the team expanded. 

Introducing Engineering Managers and Leads 

To manage the growing team, we introduced a layer of engineering managers and leads. These leaders were responsible for the day-to-day management of engineers, helping with everything from setting individual goals to removing roadblocks. This allowed the broader engineering leadership to focus on higher-level strategic decisions while ensuring that every engineer had the support they needed to succeed. 

Regular Feedback and Reviews 

We fueled the success of our scaling structure with the scaling structure itself. Yes, you read that right, and no, we do not claim to have invented a perpetual motion structure, I will explain. We implemented regular feedback loops in each and every single node of our structures, spanning from interviews through onboarding to exit interviews and everything in between. Also, we ensure feedback and metrics are collected, analyzed and fed back into the system to ensure everyone in the team had a clear understanding of how we were performing and where we could improve. We encouraged a culture of open, constructive feedback.  Not only did it help the engineers grow, but it also ensured that we were intentionally improving our structure and the team as a whole. 

7. The Challenges of Scaling 

Scaling is never without its challenges. Throughout this journey, we faced several hurdles that required careful thought and strategy to overcome. 

Onboarding New Engineers 

Bringing in new engineers quickly can overwhelm existing teams if not done correctly. We invested heavily in our onboarding process (we will talk about this in a later post), ensuring that new hires have a smooth transition into the company. We created detailed onboarding documents, set up mentoring programs, and scheduled regular check-ins to ensure that new engineers were integrating well into their teams. 

Communication and Alignment 

Another challenge we faced was ensuring that communication remained effective as the team grew. With a small team, it’s easy to keep everyone on the same page, but as the team expanded, we had to be more intentional about communication. We implemented regular all-hands meetings, team stand-ups, and project syncs to keep everyone aligned on goals and priorities. 

Maintaining Quality 

One of the most significant challenges in scaling is maintaining the quality of the work being produced. As we brought in more engineers, we needed to ensure that they were all producing high-quality code and following best practices. We did this by implementing rigorous code reviews, automated testing, and continuous integration pipelines. These processes allowed us to catch issues early and maintain the quality of our product, even as the number of contributors grew. 

8. Looking Ahead: The Next Phase of Growth 

As we reflect on our journey from 16 engineers to over 100, it’s clear that we’ve come a long way—but our growth story is just starting. Tech space continues to evolve rapidly, and as a company with a very ambitious mission, we need to stay ahead of the curve. 

Stay tuned—we’ll be talking about scaling from 100 to 1,000 engineers in less than a year before you know it!  

Looking ahead, we plan to continue scaling our team with an even greater focus on diversity and inclusion. We believe that. 

Join us on this exciting journey!