AI platforms
Trend 14: Integrated AI lifecycle tools to drive industrialized AI
Enterprises cannot afford to take an artisan approach to AI and experiment with pilots and a handful of disparate AI systems built in silos. Without focusing on achieving AI at scale, data scientists created “shadow” IT environments on their laptops, using their preferred tools to fashion custom models from scratch and prepare data differently for each model.
The AI lifecycle involves various stages, from data collection, data analysis, feature engineering, and algorithm selection to model building, tuning, testing, deployment, management, monitoring, and feedback loops.
Infosys proposes a simplified ML to prepare data, develop the data model, deploy, and optimize models in production.
Optimizing AI model performance is an emerging area in the AI lifecycle. Intel provides toolsets to introspect model and measure performance, optimize math libraries and algorithms to improve model performance, and optimize versions of popular serving engines.
Based on client interactions, we are seeing a rise in the adoption of end-to-end AI lifecycle development tools, including H2O.ai, Kubeflow, and MLflow. However, the standardization of these tools and pipelines is still under progress.
A U.S.-based telecom player struggled with AI development, as it used multiple tools and had limited data for model development. The company partnered with Infosys to develop an industrialized AI system to build a pipeline for AI and ML developments over AWS. The telco was able to improve returns on its AI and ML investments.
AI platforms
Trend 15: From data scientist to data engineer with automated ML
Data scientists spend around 80% of their efforts on finding data rather than building AI models. Creating an AI model from scratch needs effort and investment for collecting datasets, labeling data, choosing algorithms, defining network architecture, establishing hyperparameters, etc. Further, the choice of language, frameworks, libraries, client preferences, etc., differs from one AI problem to another.
Specialized roles such as data engineer and ML engineer offer skills vital for achieving scale. With the help of a rapidly expanding stack of technologies and services, teams have moved from a manual and development-focused approach to an automated, modular, and fit-to-address approach, from managing incoming data to monitoring and fixing live applications.
A global investment firm's data scientists were struggling to get the right data experience. The firm, in partnership with Infosys, built an efficient solution to plan, prepare, and predict data. The solution reduced the time spent on identifying the right tools and data, boosting data scientists' efficiency by up to 65%.
