Farooq Shaikh
Nominated Award: Young AI Role Model of the Year
Linkedin profile of Person: https://www.linkedin.com/in/farooqas/
Sonalake is a software partnering company that helps enterprises to innovate using software. We reinforce existing know-how and data with cutting-edge analytics, automation and visualisation technologies.
Sonalake has deep domain expertise in telecommunications. Over nearly 20 years, Sonalake and its predecessor organisations have delivered over 400 projects to large and small network operators across the globe. We have deployed innovative web-scale telecom management solutions globally with clients such as Vodafone, BT, eir, enet and Three.
One area of particular focus for us is Network Service Analytics. In 2016, we established a dedicated Analytics Practice within Sonalake to consolidate and further develop our expertise in the area.
The Analytics Practice Team grew from just eight people to 40+ specialists and provides strategic advice, systems integration and product solutions to operators and vendors, addressing the most demanding challenges facing the industry.
The team helps enterprise clients analyse how their underlying data can be translated to provide insights and democratised to provide access across relevant user groups.
The team excels in three main areas:
Analytics Applications
Custom development of applications that cannot be fulfilled by available COTS products, including embedding analytics modules within vendors’ software products.
Data Visualisation
Advisory services on the design and presentation of analytics outputs – primarily dashboarding and reporting – based upon expertise in user experience (UX) and the available tooling, including commercial BI and open source software products.
Data Science
Advisory and development services for advanced analytics features based upon contemporary machine learning and data science capabilities.
The team has experienced lots of successful customer projects across market verticals and its efforts were recognised for the impact of their work in transforming data and visual analytics in the telecoms vertical. These include a major Data Analytics Enterprise Platform for Travelport, and a Data Visualisation and Executive Dashboarding solution that has been adopted by several large brands in the telecommunications vertical.
Reason for Nomination
Since joining the Sonalake data analytics team in April 2021, Farooq has proven to be a great asset to the team and the wider company. He has proven himself an innovator in the growing AIOps space and how it relates to our VisiMetrix network service analytics product.
Two projects, in particular, stand out.
The first speaks to Farooq’s general approach to the AI and machine learning community.
Tasked with a time series forecasting problem in the telecom domain, Farooq set about comparing a variety of forecasting methods. These ranged from traditional ARIMA and SARIMA techniques to more modern approaches such as Prophet and LSTM. There was no clear winner – each suffered from limitations in specific scenarios.
Recognising that accuracy was vital, he sought out alternatives and identified a potential solution in multi-layer metamodelling. Farooq researched several possible approaches before landing on one that met our needs.
Applying a Prophet model to the raw data and feeding the output to an LSTM autoencoder for multi-step forecasts led to a 13% improvement in forecast accuracy. In addition, this approach came with the advantage that it could factor in external data that might influence forecasts.
Based on this work, Farooq presented his findings at the PyDataGlobal conference in July 2021. The second project also involved forecasting. Farooq identified the value in understanding how weather conditions affect quality of service in mobile communication networks. Farooq faced technical challenges souring localised weather data and worked with domain experts to understand and identify metrics and key performance indicators.
The brief was also complex, but Farooq’s tenacity paid off, and we could forecast KPIs based on their historical performance and weather forecast data.
While primarily research-led, this project uncovered some fascinating real-world conclusions. First, the weather does indeed have an impact on the network quality. Second, the model could predict future network quality with high accuracy, though the accuracy of weather forecast models limits the prediction horizon.
In 2022, the Analytics Institute of Ireland shortlisted Farooq’s work in the analytics innovation category.
In addition to the above, Farooq regularly represents Sonalake at industry events to showcase our expertise in the field of AIOps. He has mastered his brief and networked with counterparts within the industry and grown awareness of Sonalake.
Finally, Farooq has also taken responsibility for mentoring students hired as part of our graduate programme. In addition, he has run knowledge-sharing workshops within Sonalake to help software engineers elsewhere in the organisation get a foothold in AI.
We are hugely proud of Farooq and his contribution to date. We often have to remind ourselves that Farooq has only been with us for a relatively short time and is only at the beginning of his career journey. He has a bright future as a leader in our industry.
Additional Information:
In addition to the above, Farooq has continued to develop his own skills, obtaining MLOps certification and commencing his AWS certification path.
Supporting Links:
• Building a Meta Forecasting Model with Prophet and LSTM for Time series Forecasting | PyData Global 2021 (https://pydata.org/global2021/schedule/presentation/179/building-a-meta-
forecasting-model-with-prophet-and-lstm-for-time-series-forecasting/ )
• Farooq Shaikh | PyData Global 2021
(https://pydata.org/global2021/speaker/profile/258/farooq-shaikh/ )