Published on: Oct. 5, 2023, 10:06 a.m.
Embarking on a journey back in time, I find myself reminiscing about a project that not only shaped my consultancy career but also brought innovative solutions in the realm of Artificial Intelligence (AI) and Machine Learning (ML). Today, I'm peeling back the curtain (as much as I can, considering the NDA) to share my experience from August 2015 to May 2017.
The inception of this journey was as unexpected as it was exciting. A message on LinkedIn from an unfamiliar face sparked a collaboration that would immerse me into extensive R&D in ML, aligning seamlessly with my ongoing 2nd PhD in the field. The project? To devise an intricate call center scheduling algorithm, accommodating over 1000 employees, navigating through various time zones, languages, and regulatory requirements e.g. working hours, shifts etc.
My role required deep dive into research, seeking an optimal AI software solution that would not only meet the client’s immediate requirements but also cater to future needs, with a preference towards open-source solutions. A meticulous period of discussions and brainstorming sessions with the client helped carve out the precise requirements for the algorithm. This was followed by an exhaustive literature review and analysis of prevalent solutions, both proprietary and open-source, culminating in a comprehensive report detailing the most viable solutions, each accompanied by their respective pros and cons.
This entailed:
The client, thrilled with the insights and recommendations, entrusted me with the next step: developing a Minimum Viable Product (MVP) of the AI algorithm and assembling a team to bring it to life. I onboarded three of my brightest final-year students, and together, we embarked on a development journey from December 2015 to May 2017, successfully implementing multiple enhancements to the original AI algorithm, which was eventually deployed in June 2016.
The subsequent phase involved:
Ensuring the project’s success required meticulous planning and resource management, involving:
The algorithm was not only successfully implemented in production by June 2016 but also underwent numerous enhancements, proving its robustness and efficiency.
I also advised the company on software design, advocating for a modular approach to facilitate seamless algorithm updates without disrupting the entire software stack. The AI optimization algorithm was embedded with an internal API, accessible organization-wide, and was supported by parallel-running modules for production, backup, testing, and development.
In conclusion, this project was not only a successful endeavor in delivering the requested algorithm (and more) but also a tremendous learning experience for me, particularly in contract negotiation, management, strategy, recruitment, working in large international team, remote work, Agile, etc.
We were remote first, even before remote existed. I funded the company bitt.solutions and was the CEO and CTO.
While the NDA restricts me from divulging specifics, I can share that the client had a turnover exceeding $100M and was acquired by a big company (with a turnover of $1-2B) between in 2018. I am inclined to believe that our AI algorithm played an important role in this acquisition.
If you need a consultant or partner company to help you achieve your dreams do not hesitate to contact me.