Now the software decides: inter-district transfer of ZP teachers underway in Maharashtra

The highly anticipated inter-district automated transfers of teachers from zilla parishad schools in Maharashtra are underway, officials said, adding that teachers wishing to transfer to other districts have started indicating their choices.

“Inter-district transfers of teachers to ZP schools started from midnight between August 19 and 20,” zilla parishad CEO Ayush Prasad told the Indian Express on Saturday. For the first time, decision-making has been entrusted to software based on a fixed set of rules, he added.

The software, Prasad said, initiated the transfer process following all the rules in accordance with the government resolution of April 7, 2021. “The software is fully automated without any possibility of human intervention. The entire log of every rule applied and every decision made is retained and searchable,” he said.

Automated decision making

Officials said the software can run up to 34 loops. In the simplest loop, for example, a teacher from Pune would transfer to Satara and a teacher from Satara to Pune. In a loop of three districts, a teacher from Pune will go to Satara, a teacher from Satara will go to Solapur and the teacher from Solapur will transfer to Pune – thus filling mutual vacancies. “The most complex operation that would be performed by the software would be to create a loop of the 34 zilla parishads in the state,” Prasad said.

“This is a solution that uses parallel computing blockchain and hyperprocessing to compute ‘n high to n’ permutations and combinations to produce the most optimal computed results on very complex data and processing several rules without a single human intervention”, explained Nilesh Devidas. , CTO of Vinsys, the company contracted to develop and maintain the software.

“It is an example of artificial intelligence. Most of the software used in the government is used to collect data, populate dashboards, retrieve stored data for the purpose of facilitating decision making by the concerned official. Pune zilla parishad has already used automation to check eligibility in Mahalabharthi Software. I had the difficult task of explaining to very influential people why I cannot help with the transfer of their relative, because the system is completely automated and governed by rules,” he said.

Intensive computing power

For cross-district forwarding, six nodes consisting of eight virtual machine processors in each of the nodes, hosted on a cloud, are used to compute data in 14 rounds. Computing the data is expected to take 31 hours. A blockchain was created to improve computing power, Prasad said. “Simply put, 48 computers would calculate continuously for 31 hours to process the data. Each calculation and each decision of each of the computers is recorded in the logbook.

Data Accuracy

Numerous administrative measures have also been taken, such as the approval by the divisional commissioners of the lists of caste categories in each group of teachers in each district, officials said. An online social audit of the data entered by the teachers was carried out. Full transparency has helped the system gain teachers’ trust, Prasad said. The CEO of ZP said the teacher transfer policy was adopted after extensive consultations with teachers and their unions. “Social media has been used to generate opinion and facilitate debate on several contentious issues. The policy has been wholeheartedly accepted by all sections of teachers,” he added.

Security audit

Prasad said a Cert-IN security certificate, Standardization Testing and Quality Certification (STQC) compliance, internal compliance to maintain the software’s Sigma-6 rating, App Scan and other certificates of security were taken. “The cloud provider is authorized to provide services to banks and provides these services to multiple banks and multiple Union and State governments,” he pointed out.

Audit trail

All data is stored in an audit trail with an unmodified blockchain model. Any changes to data require OTP authentication from the respective data owner, officials added.