• Home
• /
• Tech

Smart machines are ready to take over

The right technology stack to deploy models in production go a long way in the success of an AI-driven organisation

Artificial Intelligence (AI) is a field that focuses on the creation of smart machines or processes that mimic human actions. Machine learning, computer vision, speech recognition, image recognition, among others, are considered different branches of AI. A lot of recent technological advances including self-driven cars, Siri and Alexa are made possible through AI. In healthcare, AI has been playing a vital role in drug discovery and disease prediction. AI brings in value for organisation and businesses where a lot of human actions are mechanistic in nature and can be well characterised by data. AI is transforming the world through automation of several tasks to reduce turnaround times as well as provide personalisation to increase the quality of product and services. 

Data is the oil that drives AI. AI relies very heavily on data covering all possible dimensions. A data-driven culture is required to nurture AI in an organisation. Organisations would go a long way with AI by having the right approach to gathering and maintaining data. Lot of organisations fail in AI due to lack of data vision. AI methodologies would differ based on the type of data being analysed. Structured data have a set of methods whereas unstructured data namely speech, text and images require certain additional methods to process them.
Machine learning is another core part of AI. Machine learning modelling process encompasses exploratory data analysis, feature engineering, and model building, validation and deployment. Learning without any kind of supervision requires an ability to identify patterns in streams of inputs, whereas learning with supervision involves classification and regression. Deep learning, a sensation in AI in recent times, takes feature engineering away and provides superior performance compared to traditional models.
Often overlooked piece in AI life cycle is the deployment of AI models. Having the right technology stack to deploy models in production would go a long way in success of an AI-driven organisation. A long-term vision is required in order to ensure tech stack is stable enough to accommodate futuristic models.
Open source tools have supported the exponential growth of AI applications. With Python/R/Scala/Julia as the major tools, AI has moved from being a research topic to realising its potential in real world applications.
- Vineet Shukla is Director of Engineering - Data Science and Machine Learning, United Health Group, India and Ravishankar Rajagopala is Engineering Manager - Data Sciences, United Health Group, India.