When it comes to Insurance, prolonged delays in processing data and the need for human involvement in extracting and confirming information have posed significant obstacles since time immemorial. These manual procedures not only demand a hefty resource stack but are also error-prone, resulting in bottlenecks and operational inefficiencies. That is when Kamlesh Goyal’s full-stack insurance company Digit General Insurance came into the picture to address this challenge. Digit Insurance’s robust data platform enables real-time data processing that aids in decision-making, risk assessment, fraud detection, and customer engagement.
“Technology is the backbone of Digit and over the years we have embedded various tech innovations into the systems that have aided in a complete overhaul of how insurance products are experienced by Indian customers today” Vishal Shah, head of data Science, Digit General Insurance told AIM.
And to make this possible, Digit Insurance boasts a big AI and analytics team with over 130 employees distributed across distinct divisions such as analytics, data science, business analysis, and API Integration.
Based in Bengaluru, the general insurance firm achieved unicorn status in 2021, a mere five years following its launch, thanks to a $255 million funding round led by General Atlantic and Multiples Private Equity. It is also backed by the likes of Sequoia Capital India, Fairfax Group and TVS Capital Funds.
AIM got in touch with Shah and Amrit Jaidka Arora, chief of human resources, Digit to understand their data science applications, hiring process, work culture and more.
Currently Hiring
Digit Insurance is expanding its data science team and is looking to hire AI/ML engineers and BOT developers. They are also looking for data engineers (data warehousing and data lake teams) as well as visualisation developers. Furthermore, they are actively seeking individuals with a robust understanding or a strong passion for insurance and API Integration to join their team in a Business Analyst + Technology-oriented role.
Inside Digit’s AI & Analytics Lab
The organisation leverages AI and ML extensively in its operations, with a focus on its robust data platform that securely stores diverse structured and unstructured data. This platform integrates data from various sources, like customer interactions and transaction records, for analysis, particularly crucial for real-time applications like fraud detection. Data engineers play a pivotal role, in ingesting and transforming data through ETL processes, ensuring quality, and optimising query performance.
“The team also uses NLP to streamline document processing, reducing turnaround times and aiding in fraud detection. Their AI-powered chatbots reduce query resolution times, while computer vision expedites pre-inspection for vehicles and enhances four-wheeler claims assessment accuracy,” added Shah.
They also harness its extensive datalake for ML algorithms, driving both one-time analysis and real-time solutions based on insights derived from the data.
Tech Stack
The team primarily uses SQL, Postgres, No-SQL database, Python either through Jupyter Notebook or Visual Studio. Various frameworks like TensorFlow, PyTorch, Hugging Face, NLTK, spaCy, are also used in our processes.
Transformer or GPT models are often employed for natural language processing tasks like chatbots or sentiment analysis.
“When it comes to generative AI, we have initiated pilots on various use cases including chatbot, help desk and more. We are also evaluating various LLM for their efficacy in providing the results,” commented Shah.
Addressing Ethical AI
The adoption of AI in the insurance and financial sector presents significant ethical and regulatory challenges.
According to Shah, the key concern revolves around safeguarding user data privacy and confidentiality, especially when extracting information from customer documents. Compliance requirements necessitate transparent data handling and storage practices, including clear disclosure of data processing and usage procedures.
Additionally, mitigating model bias is crucial to prevent unfair or discriminatory outcomes. Complying with regulations, particularly in sectors like insurance, demands transparency in decision-making processes, compelling organizations to articulate how their AI models arrive at decisions and predictions to maintain compliance. Addressing these challenges involves robust data protection measures, bias mitigation strategies, and transparent model explanation practices.
Interview Process
“The interview process at Digit begins with an initial round of discussions, either conducted internally or outsourced, aimed at evaluating the candidate’s qualifications. Following this, candidates are given a case study or assignment that assesses their core skills. The final technical discussion is built upon the results of this assessment. Lastly, a third and final round takes place, during which HR conducts an offer discussion with the candidates.
Digit’s hiring process for data science roles involves a focus on four distinct tech sub-departments: Data Science (encompassing AI, ML, and BOT), Analytics (involving Data Engineering and Data Visualization), API Integration, and Business Analysis. For these roles, we seek candidates with a diverse skill set that includes proficiency in Python, SQL, PL, SQL, Qlik Sense, NLP, and Computer Vision.
But what is the common mistake that candidates often make?
“Candidates often miss reading or understanding the JD in detail and apply for roles that are not similar to the profile mentioned. Though their CV would mention some of the skills, if they have not worked in-depth on some of the core skills we are looking at, clearing the interview and technical rounds can be tough,” Amrit Jaidka Arora, CHRO, Digit General Insurance told AIM.
Expectations
When joining the data science team at Digit, candidates are introduced company culture that is both flexible and transparent, providing a plethora of opportunities for growth and development.
“We are known for our diversity, offering the chance to collaborate on significant, scalable projects and engage in cutting-edge initiatives at the forefront of data-driven decision-making, thereby gaining valuable experience in emerging technologies,” said Arora.
In return, Digit expects candidates to possess strong coding skills, a robust logical and analytical mindset, and relevant experience aligned with the job descriptions. Candidates should also be prepared to thrive in an agile and dynamic work environment.
Work Culture
“In 2017, we started as the 27th General Insurance Company, knowing the tough competition and industry norms but we saw it as a chance to change the insurance sector. Our approach focused on creating a lasting culture that matches our goal of simplifying insurance. We value challenging conventions and transparency, promoting a positive workplace,” said Arora.
Digit aims to create an open culture based on four key principles. The first is ownership, where employees adopt an owner mindset and take responsibility for addressing feedback and collaborating to improve. Second is people relations, emphasising treating everyone with respect and empathy. The third principle is ‘evolve’, highlighting the importance of continuous growth, learning, and mutual support in achieving their mission. Lastly, they embrace a ‘no hierarchy philosophy’, promoting accessibility and equality among all employees, fostering a flat organisational structure to encourage simplicity and inclusivity.
In terms of perks and benefits, Digit offers a comprehensive package to its employees, which includes ESOP, various insurance coverage, attendance bonuses for specific departments, health and wellness benefits, relocation support, study assistance programs, car lease support, creche facilities, and referral bonuses and many more.
“So if have the skills and drive to deliver rapid, innovative, and intelligent solutions while efficiently collaborating with colleagues and customers with cognitive solutions, Digit Insurance is your ideal for success,” concluded Arora.
Read more: Data Science Hiring Process at Dream11
The post Data Science Hiring Process at Digit Insurance appeared first on Analytics India Magazine.
