A team of Google researchers has created an AI model that designs next-gen chips faster than human designers. The model could build a design that optimises the placement of components on a chip within six hours. The team used a dataset of 10,000 chip layouts for an ML model and trained it using RL. The new AI-generated layouts were used in Google’s latest AI accelerator (TPU-v5). AI uses past experience to become better and faster.
Google cloud launched three new ML-based services that will provide real-time data insights to users. The three services: Datastream allows users to replicate databases in real-time. Analytics hub lets users curate, create, and manage analytics exchanges securely and in real-time. Dataplex is a combination of Google cloud and open-source that allows users’ analyze their data at scale. A survey by Gartner found the average cost of poor data quality at $12.8mn per year.
Twitter has launched an initiative on “responsible ML ” to reviews its algorithmic fairness with the aim to bring transparency on how it tackles “the potentially harmful effects of algorithmic decisions.” Concerns on algorithms used by internet platforms, which some allege promote violence, hate, racial/gender bias, is at an all-time high. Recently, controversies at Google’s AI ethics team resulted in the firing of two researchers and the resignation of a scientist.
60% of Haryana-based start-ups may not invest in the state following the law which stipulates 75% job reservation for locals with a monthly salary of up to Rs 50,000 in the private sector. Nasscom, in a survey, said that the law which comes into effect from May 1, is going to hit hard to start-ups as 67.3% of their workforce earn below Rs 50,000 per month. Besides, startups involved in deeptech, edtech need high-end AI and ML skill resources, which is hard to find locally.
V. Padmanabhan, VP – Engineering, GlobalLogic led team has created AI and ML tools to look for candidates certifications, present and past companies, their research papers to create a summary of thousands of resume just through systems. A team of 10 professionals took at least a few weeks to look at 1000 resumes, do talent identification for fitment into the roles. In the new automated system, the efforts have come down to a day or two.
Shree Cement during the expansion of the business had few priorities – Production capacity, maximum hardware utilisation and reduction in the costs of a database software license. It switched to new hardware and shifted the workload to IBM POWER9 and introduced IoT integrated apps for supply chain management along with ERP solutions. For predictive analysis of plant maintenance system and visual analytics-based app, it will introduce AI and ML.
InCred Finance, a Mumbai based financial company has appointed Philipp Orgler, who holds 20 years of experience in financial analytics, for developing AI and ML tools. Orgler has headed the dept of structured products at Barclays Capital, ran structured notes business at Mizuho, worked with Deutsche Bank and done analytics work for Rokos. He is masters in Theoretical Physics from Karl-Franzens University and did research in Mathematical Physics at the University of Tokyo after graduation.
Golang, C# (C-Sharp), specialisation in Hadoop and ML jobs are the top-paying profiles earning Rs 24-35 lakh per annum, as per a Randstad study. These jobs saw an avg rise of 25-30% in salaries between 2019 and 2020. Bengaluru is the highest paying metropolis across the junior, mid and senior-levels with an avg salary of 16-20% higher than peers. IT and Internet and e-commerce top the list in terms of the industry pay package.