During a technical discussion, one of our esteemed customers in the rail domain, highlighted the problems faced in deploying AI models and the challenges associated with identifying the right kind of MLOps tools. I could understand and visualize the use case and wanted to replicate a similar scenario in-house as it could be a first step in reproducing the challenges faced by the customer.

Moreover, it could be demonstrated during the planned workshop at the customer site. The principal stakeholders from Rail Practice agreed to this idea. Thus, I embarked on an ambitious journey with a goal to develop a prototype of an end-to-end pipeline which included AI model building, deployment and monitoring. The vision was clear, but the path ahead was uncharted. Being part of CoE, I did not have a team – and I needed a team of exceptional individuals, driven not by mere obligation but by an insatiable passion for engineering, a commitment to continuous learning, and an innate collaborative spirit.

I initiated a discussion with a leader of Application Engineering CoC of Product Engineering Practice and explained the requirements. He was confident that it could be done. We then formed a dedicated team of skilled engineers, including an AI specialist, to delve into AI and MLOps intricacies. They were not directly reporting to me, yet they embraced my project as their own, with unwavering dedication.

Like a well-rehearsed orchestra, the team harmonized their skills, each member playing a pivotal role. We finalized anomaly detection using deep neural networks as the use case to be demonstrated and started regular rhythm meetings. The team meticulously architected the solution, designed the intricate data pipelines, selected the right MLOps and DevOps tools and expertly integrated the AI models into the application, ensuring seamless deployment and monitoring – thereby bringing my dream to life.

Collaboration was the team’s hallmark and they readily shared knowledge. Ideas flowed freely, each suggestion carefully considered, and each challenge tackled with collective ingenuity. Amidst all the turmoil of development, they never lost sight of the customer’s needs. They appreciated feedback and ensured that the prototype was aligned with the customer’s roadmap.

Within the expected target time of four weeks, the prototype was complete – a testament to the team’s extraordinary dedication and commitment. The MLOps pipeline worked correctly, and the lifecycle of the AI model could be demonstrated without any glitches. The whole team joined online, when I presented the MLOps session during the onsite customer workshop in Paris. They demonstrated the full functionality of the system we developed, answering all the queries from the customer.

Waves of gratitude arise in my heart as I think of this journey which was entirely built on our organization’s culture pillars of collaboration, customer focus, aspiration and continuous learning. It was indeed a collaboration between industrial and digital software teams. This team of exceptional engineers ignited a spark, illuminating the path towards more opportunities.

Let me conclude with this note:

“To this remarkable team – Your contributions have been invaluable, and your impact will be felt far beyond the confines of this project. You have set a shining example of what can be achieved when passion, dedication, and collaboration converge. Thank you for believing in my vision and for helping me bring it to life.

Written by Sindhu Ramachandran S

on 21 Dec 2023

CoE Leader for Artificial Intelligence,

Quest Global