For some time now, iMining Technologies, an investment company has been focused on acquiring and developing of future technologies. However, with the growth of artificial intelligence (AI) and generative AI (gen AI), the team made a strategic decision to start a new subsidiary that would focus on AI – AiMining Technologies.
As a subsidiary was borne from that decision, the company has the mandate to build on three strategic pillars: AI core applications and technologies; AI infrastructure; and semiconductors.
The value proposition of building on these three pillars is the creation of not just a product but an ecosystem of talent, jobs, IP and AI sovereignty. Headquartered out of Vancouver, Canada, the company takes a multidisciplinary and applied research-based approach to problem-solving.
AIMining’s focus is on research and AI sovereignty, Baig believes this makes them different by design. Working in the Gulf region with their head office in Canada makes for an interesting mix.
“We’re approaching this market uniquely: Building the Gulf’s own fine-tuned LLM and the first AI supercomputer and, more importantly, reinvesting into diversified talent across the region. We are concurrently planning to build Canada’s first indigenous languages LLM (Cree, Metis, Inuit and more). To us, being part of the region, building talent, and making sure we do our part for AI sovereignly matters. That’s part of who we are and what makes us different, and we are proud of it,” said Mirza Mushtaq Baig, Chief AI Officer (CAIO), AiMining Technologies, in an interaction with edge/.
The team believes currently, the Middle East market is witnessing significant growth and interest in adopting AI and advanced technologies across sectors such as finance, healthcare, government services, transportation, and education.
Governments and private sectors are investing in AI research and innovation, leading to diverse applications such as smart cities, healthcare solutions, and fintech innovations.
“In the UAE, everything from advanced quantum computing to AI to post-quantum cryptography is being done. In KSA, due to investments by powerhouses like Amazon, Dell, IBM and ServiceNow, there is immense potential to become leaders in the AI space,” said Baig.
The workings
With a team of globally renowned AI and ML engineers, data scientists, software developers, post-doc research fellows, and a distributed network of advisors, the team analyses client requirements, collects and processes data, develops AI models, trains and evaluates them, deploys solutions, and provide monitoring and maintenance for optimal functionality.
The goal is to leverage AI to solve complex real-world problems, drive innovation, and deliver value to clients across industries, including education, mining, healthcare, cyber defence, and food and beverage.
“We’re not building in a vacuum or isolation; we live in the real world, so we are creating real-world jobs, working with the best researchers, partnering with the best universities, and creating and delivering real-world value. To us, AI is not some black-box proprietary or open-source (take your pick) brain teaser or academic exercise. When built responsibly, AI is all about creating jobs, developing and nurturing talent, and solving complex and real-world problems like climate change,” said Baig.
The core technology of the product was developed through a systematic process involving research, conceptualisation, prototyping, development, integration, testing, and deployment.
Building on the MVP
The initial steps in building the first product involved market research, defining core features, developing a minimum viable product (MVP), and iterating based on user feedback. The product subsequently evolved through continuous improvement, iterative development cycles, testing, and user feedback integration.
Updates were made to enhance features, optimise performance, and improve user experience, leading to scalability, seamless integration with third-party services, and effective monitoring through analytics. Strategic adaptations were made to stay aligned with market changes and ensure the product remained competitive and valuable to users and stakeholders.
The team is building a suite of products with the support of the NVIDIA Inception programme and Microsoft Founders Programme.
Building the prototype
The team begins with thorough research to understand the problem and then conceptualise product functionalities. Prototyping and iterative development help validate ideas and gather feedback, leading to the creation of core technology components, such as algorithms and data processing pipelines.
Rigorous testing and quality assurance ensure a seamless user experience while continuous monitoring post-deployment allows improvements and updates in the future.
“As a multifaceted venture, we are building open-source custom-built LLMs primarily for public sector companies through a MoE approach and RAFT (retrieval augmented fine tuning). The great thing about working with the public sector is its openness to consultations, continuous feedback, and bringing various stakeholders to the table,” added Baig.
He explained being ardent believers in representation, the team has built the core product with a multi-stakeholder opinion at the core of their build. They have also used differential privacy in data and machine learning and embraced zero-trust principles.
“Such practices are essential when you’re building something as complex as an LLM or a smart AI solution for healthcare (such as in a genome project). You must ensure that privacy, data protection, and data provenance are followed. On top of that, we use standards to govern our builds: ISO 42001, NIST, and the new IEEE global standard,” said Baig.
The product development process has several moving parts, including algorithm complexities, data quality assurance, system integration, security, scalability, performance, and user experience.
“We must reckon with those by breaking down algorithms, optimizing code, collaborating for data quality, developing robust APIs, implementing security measures, and prioritizing user feedback for iterative improvements. These efforts ensure the successful development of a scalable, secure, and user-friendly product that meets the needs of both users and clients,” said Baig.
The core technologies used in product development vary based on specific needs but typically include programming languages like Python and Java, frameworks like React and Django, cloud services like AWS and Azure, databases, and DevOps tools.
Continuous evolution
To ensure continuous evolution and growth, organisations should stay updated with technology trends, gather and act on user feedback through iterative development, embrace agile practices, foster cross-functional collaboration, encourage continuous learning, prioritise experimentation and innovation, monitor and optimise performance, and stay connected with market trends and customer needs for ongoing improvement and expansion.
The team believes AI technology is expected to advance significantly with improvements in ML algorithms, natural language understanding, and computer vision. So, they will be focusing on AI ethics, responsible AI practices, and regulatory frameworks. The integration of AI with robotics, AI-driven personalisation, edge computing, and AI for sustainability, too, are key areas of development.
“Merging into quantum and post-quantum computing will also become increasingly important. It’s just a matter of time before quantum computing comes to the fore, so we might as well get ahead of the curve with PQC now. And from there, with the field of AI changing rapidly, we have a lot more in store for our clients,” said Baig.
