India is banking on electric scooters like never before. As per Niti Aayog, the electric two-wheeler industry is expected to scale exponentially, taking over 80 per cent of the market by 2030.
Leading the EV revolution is Bengaluru-based Ather Energy. The company witnessed a 1,149 per cent jump in sales after its expansion across 13 cities across the country, where it sold close to 1,799 units in July 2021, compared to 144 units sold in the same month last year, as per FADA.
But, there is more to Ather Energy’s success story. The Ather ecosystem collects a lot of raw data. This includes all the data collected around the scooters, the grid — fast-charging infrastructure, pre-sales interactions, post-sales service, social media interactions, and everything in between; thus, putting Ather Energy in a unique position over other players in the market.
AIM Daily XO
Join our editors every weekday evening as they steer you through the most significant news of the day, introduce you to fresh perspectives, and provide unexpected moments of joy
With its objective of the intelligence platform, the company believes in leading from the front in terms of the overall two-wheeler experience. Its key objectives include:
- Leveraging the data collected to provide delightful, smart features to its users that are hard to replicate. Some of the recent OTA (over-the-air) updates built and deployed by Ather include automatic indicator and turn detection, theft tow detection, range prediction, etc.
- Enabling the organisation to move from a reactive to a proactive way of operating for better decision-making by developing predictive models for diagnostics, marketing, sales conversion, and churn.
For instance, Ather Energy’s automatic indicator-off feature detects a turn in the vehicle and triggers turn-off. Unlike other two-wheelers or cars, the feature is not based on the turning of the steering or handlebar and works off the patterns of a rider as they ride and the changing angles of the scooter. Ather has built the feature to detect even lane changes and U-turns and regular rights or lefts.
Download our Mobile App
“There is a significant signal-noise elimination that needs to be done before having the model learn the patterns to be able to predict a turn to have occurred/completed. We stream high-frequency data and detect these turns at a latency of less than a few milliseconds,” shared the Ather Energy team.
The company is currently hiring vehicle data scientists with three to five years of experience. A candidate who is enthusiastic about IoT and wants to build intelligent, scalable products with embedded intelligence would be perfect for this job role.
Ather Energy has about 10 data scientists who work with edge devices and cloud development, feature deployment, and analysis of the vast data it collects. The company’s data science team is divided into two streams — vehicle intelligence and web intelligence.
Ather Energy’s interview process is pretty straightforward. “While we do expect and assess a traditional understanding of AI/ML in our technical interviews, we also look for ease of adaptability to switching domains. We give our candidates some insight into the kind of problems we cater to, with a use case round where a contextual Ather dataset is provided with the problem statement. To evaluate general aptitude and application of analytical thinking, we use this problem to delve deeper into applications at Ather Energy,” said Sunitha Lal, CHRO at Ather Energy.
“We have interviewed experienced candidates who have years of experience using Big data technologies but have traditionally worked more on sub-optimal model development,” said Lal. She revealed that Ather primarily focuses on strong ML, CS, and statistical basics and looks for aptitude in candidates to learn a new domain and apply traditional statistical techniques in this space.
At Ather Energy, the data science teams solve a wide gamut of problems. The changes are rapid. The company looks for candidates who can help hypothesise quickly, build iterative solutions and adapt to various domains and tech stacks to address the rapidly evolving needs of the organisation. For instance, the team supports various business needs outside of the vehicle as well. Some of them include:
- Smart vehicle: Work with the embedded edge device to build ride critical algorithms while simultaneously working with the cloud infrastructure to query and monitor data and metrics to tune the algorithms (safety and ADAS features).
- Mobile app: Build and measure unique features with well verified A/B tests for performance benchmarking, build smarter personalised features that enhance customer experience.
- Business and marketing: Identify churn of their premium subscription customers, support marketing and social media analytics to optimise their stories and pitches.
- HR: Support data-driven benchmarking and performance across the organisation, streamline and automate hiring screening through text-based models to hire the best candidates.
From a career perspective, Ather Energy has multiple software platforms collecting a lot of data over the past few years. These datasets allow for faster product development and innovating products that improve user experience. But, the potential of what this kind of data can do in the future of intelligent vehicles is still largely untapped, and prospective candidates can look forward to playing a role in solving more problems efficiently using the data perspective.
As Ather Energy is scaling exponentially, the company works closely with different teams collaboratively, similar to a startup. “Things are always changing, and we follow an agile culture to adapt to these rapidly changing requirements,” added Lal.
She said most companies in this space leverage data to understand the impact of business decisions – mostly post the product development, in marketing, inventory, and supply chain. Ather Energy, however, takes data-driven decision-making several steps further.
Ather’s first vehicle was designed based on data collected using a custom DAQ on IC engine vehicles running over a few thousand collectors. From the suspensions to the ergonomics of the chassis and the seat, the algorithms used in the ECUs to the personalised range predictions, data collected across the 1000+ km built a strong real-life foundation and a stepping stone to various potential ML Models. Today, it uses ML simulations trained on fleet data to estimate the durability and reliability of parts of the vehicle.
Want to join Ather Energy?
Data science is a combination of skills, mainly statistics, programming and business application. If you are interested in applying for a data scientist’s role at Ather Energy, you should have a strong background in the above skills. Plus, the candidate should be able to conceptualise building a product from the algorithms ground up. “While complex ML models are great to build out, without productising the use-case for the need for this model, there would be no business value generated out of the effort,” said Lal.
Check out for data science job roles at Ather Energy here.