Latest news with #Bhaskaran
New Indian Express
6 days ago
- Politics
- New Indian Express
Clamour growing for union government intervention to save Nimisha Priya
PALAKKAD: Uncertainty surrounds the reported execution date of Nimisha Priya, the Malayali nurse sentenced to death in Yemen for the murder of a Yemeni national. Over the past two days, multiple media have claimed that her execution has been scheduled for July 16. However, there is no official confirmation from the Ministry of External Affairs (MEA) or the Indian Embassy so far. The initial reports were sparked by a statement from Samuel Jerome Bhaskaran, an Indian national who has lived in Yemen for over two decades and has been involved in negotiating a settlement with the victim's family through the payment of blood money. Bhaskaran told TNIE over the phone that he received a call two days ago from the chairman of the Central Prison in Sana'a, who confirmed that the execution order had been issued. Following this, Bhaskaran said the information was verified by a Yemeni national associated with the Indian Embassy in Saudi Arabia, who personally visited the prison in Sana'a and informed him of the development. 'The execution order has been confirmed. The priority now is to work with local authorities and the Indian government to seek a postponement and at least postpone the execution order of Nimisha Priya,' Bhaskaran said. Priya, a native of Palakkad, has been on death row since 2020 after a Yemeni court sentenced her for the alleged murder of Talal Abdo Mehdi, a Yemeni businessman, in July 2017. Her final appeal was dismissed by Yemen's Supreme Judicial Council in November 2023, though the court left open the possibility of a pardon if the victim's family accepted blood money under Yemeni law. Despite the claims of execution, official sources in the MEA and the Indian Embassy in Saudi Arabia have not confirmed any such development. India does not have an embassy in Yemen, and Nimisha is imprisoned in Sana'a, a region under Houthi control, making direct diplomatic intervention difficult. The ongoing civil conflict in Yemen has severely limited access for foreign missions and complicated communication. In response to the circulating news, Advocate Subhash Chandran K R, a Supreme Court lawyer and member of the Save Nimisha Priya Action Council, said the council has not received any official communication.
Time of India
07-07-2025
- General
- Time of India
EFI gets GCC nod to restore 22 ponds
Chennai: Greater Chennai Corporation (GCC) has given permission to the Environmentalist Foundation of India (EFI) to start work on restoring 22 ponds spread across zones 14 and 15. In June, EFI submitted a proposal to restore the ponds. The group proposed to deepen and widen the ponds, strengthen bunds, improve infrastructure, fence the premises, clean the waterbody of biodegradable debris, desilt the lake during the dry period, and plant native trees along the bunds to increase green cover. The project would be carried out at an estimated 2 crore. "This will prevent localized flooding. We have been a long-standing partner of GCC especially with respect to matters of water conservation and restoration," said EFI founder Arun Krishnamoorthy. Two ponds in Madampakkam, one in Allikulam, Uzhavar Keni Pond near Karapakkam, Paraserikeni Kulam, Periyakeni Kulam, Kannagi Nagar 17th Main Road Pond, Vannan Kulam, and Sathyavanimuthu Street were among waterbodies that will be restored. GCC will provide administrative support for project. You Can Also Check: Chennai AQI | Weather in Chennai | Bank Holidays in Chennai | Public Holidays in Chennai GCC chief engineer (stormwater drains) Bhaskaran said GCC has independently taken up work to restore canals within its limits. "We are increasing the concrete wall height to five feet and erecting fences to prevent the dumping of waste," he added. EFI has independently taken up bio-remediation drives to restore waterbodies across several cities. It began work on Dargah Pond in Kovalam, planting Canna Indica plants around its surface. "Stem-based plants (such as Canna Indica) technically draw nutrients from waterbodies to grow. This is beneficial as nutrient load in waterbodies often increases the accumulation of algae and brings other pollution-related hazards. This is a biological remedy to water pollution," he said.
Forbes
07-07-2025
- Business
- Forbes
Building The Backbone Of Intelligent Automation
Shinoy Vengaramkode Bhaskaran, Senior Big Data Engineering Manager, Zoom Communications Inc. As AI agents become more intelligent, autonomous and pervasive across industries—from predictive customer support to automated infrastructure management—their performance hinges on a single foundational capability: real-time, context-rich data. These agents aren't traditional analytics consumers—they're dynamic systems that require timely, reliable and actionable data streams to sense, reason and respond. What powers this intelligence isn't just the model but the pipeline behind it. The AI Agent Era And Its Data Dependency AI agents today operate in environments that demand not only accuracy but also speed and adaptability. From autonomous fraud detection in fintech to personalized content delivery in media, these agents must process data from diverse sources and react to constantly evolving inputs. Traditional extract, transform and load (ETL) systems—built for periodic batch jobs—fall short in these contexts. Instead, modern pipelines need to be event-driven, modular and responsive to real-time stimuli. And although the tools used may vary, the architectural principles remain largely consistent. Redefining Data Pipelines: From Static Flow To Intelligent Feedback Modern data pipelines for AI agents aren't just conduits for raw data—they're dynamic systems that enable feature extraction, enrichment, inference and feedback loops. These pipelines must evolve beyond rigid batch jobs toward flexible, scalable flows that support high-volume and low-latency requirements. Frameworks such as Apache Spark and Apache Flink can be considered for processing, depending on whether the task is batch-heavy or stream-driven. Spark, known for its in-memory execution, is often explored for machine learning pipelines, while Flink's fine-grained, event-time processing is suited for real-time, low-latency use cases. However, these aren't prescriptive choices—organizations may adopt other solutions based on operational constraints, data gravity or ecosystem compatibility. A High-Level Reference Architecture For AI-Ready Data Pipelines Although the specific implementation stack will vary, a high-level design often includes the following components, with several tooling options available at each stage: Organizations may opt for technologies like Apache Kafka or Amazon Kinesis to manage high-throughput event streams. These platforms are commonly used to decouple producers and consumers and allow for scalable, fault-tolerant ingestion, but are just one class of tools among several available for event-driven architectures. For distributed data transformation, tools such as Apache Spark (in batch mode) or Apache Flink (for streaming scenarios) are frequently evaluated. Both integrate well with modern data ecosystems and support flexible processing patterns. However, alternatives such as AWS Glue, Dataflow or even lightweight serverless functions may also be considered based on scale and latency requirements. Event-driven use cases—especially those requiring per-event decisions, like anomaly detection or real-time personalization—might benefit from stream processors. Flink, Kafka Streams or similar tools are often mentioned in this context, but the best choice depends on the complexity of state management, processing guarantees and integration needs. For deploying and invoking machine learning (ML) models, services such as AWS SageMaker, Kubernetes-based ML serving frameworks or even custom containerized APIs are options many enterprises explore. These systems often feed predictions back into the pipeline for real-time adaptation, closing the loop for continuous learning. Long-term and operational storage may leverage services like Amazon S3, HDFS or hybrid lakehouse architectures. These enable both analytics and machine learning workloads, but selection is typically based on governance needs, query latency and cost profiles. Workflow engines such as Apache Airflow, Argo or AWS Step Functions are often evaluated for pipeline coordination. For observability, teams may use Prometheus, CloudWatch or Grafana, although the tooling depends heavily on existing infrastructure and compliance requirements. Key Design Considerations Rather than fixating on specific tools, architecture teams should focus on four design principles: 1. Latency Tolerance: Choose stream versus batch processing based on how fast the agent must act. Low-latency use cases (e.g., trading and fraud detection) will shape your stack differently than hourly insights. 2. Scalability And Resilience: Platforms like Kubernetes or cloud-native autoscaling services provide elasticity but require thoughtful cost modeling and failover planning. 3. Modularity: AI pipelines should be loosely coupled to allow parts—such as model serving or feature engineering—to evolve independently. 4. Security And Compliance: Role-based access control, encryption in motion and at rest and audit logs should be integral from day one, regardless of stack. MLOps And Agent Evolution AI agents don't just infer—they evolve. To support this, pipelines must accommodate continuous training, automated deployments and model monitoring. Tools such as SageMaker Pipelines, Kubeflow or Spark MLlib can help with MLOps integration—but again, these are options, not universal solutions. The more autonomous the agent, the more critical the infrastructure behind it becomes. Feedback loops must be engineered into the pipeline, allowing the system to improve with each interaction and data point. A Real-World Illustration Imagine an industrial AI agent tasked with predictive maintenance. It might ingest real-time sensor data through Kafka, process signal anomalies using Flink (or an equivalent stream processor) and trigger predictions via a model hosted on Kubernetes or SageMaker. Inference results are logged to Amazon S3 or a data lakehouse for retraining. Each of these elements could be substituted based on organizational preferences, maturity and workload profiles. The Road Ahead The infrastructure behind AI agents isn't a static diagram—it's a living, evolving system. Designing effective data pipelines means embracing change, modularity and flexibility. Rather than betting on specific tools, organizations should focus on building architectures that are tool-agnostic, standards-aligned and use-case driven. That's how we future-proof AI systems—and unlock their full potential. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Indian Express
19-06-2025
- Indian Express
‘ED officials expanding their powers day by day': Madras HC says agency had no power to seal film producer Akash Bhaskaran's premises
In a setback to the Enforcement Directorate (ED), the Madras High Court has questioned the agency's authority under the Prevention of Money Laundering Act (PMLA) to seal premises that were locked at the time of a search, and reserved orders on a batch of petitions filed by Tamil film producer Akash Bhaskaran and businessman Vikram Ravindran. The ED had conducted a search and seizure operation on May 16 at multiple premises linked to Bhaskaran and Ravindran, in connection with the alleged Rs 1,000-crore TASMAC scam. The petitioners alleged that the agency sealed their office and residence illegally, even though they were not present during the raid. A Division Bench comprising Justices M S Ramesh and V Lakshminarayanan expressed concern over the ED's evolving interpretation of its powers, observing, 'Courts often remark that the PMLA (Prevention of Money Laundering Act) is an evolving legislation. But we find that it is actually the Directorate of Enforcement officials who are evolving day by day by expanding their powers.' During the hearing on Tuesday, Additional Solicitor General S V Raju, appearing for the ED, acknowledged that the agency did not have the legal power to seal a locked premise under Section 17 of the PMLA. 'The ED does not have power to seal. The lordship is right in that aspect,' Raju said. He explained that while the law allows ED officers to break open locks during a search, they chose not to in this case to avoid escalating the situation. Instead, the agency pasted notices outside the premises, stating that they 'shall not be opened' and instructing the occupants to contact the agency – an action the court said could wrongly imply that ED had the authority to prohibit entry. 'By saying ED would 'permit' them to enter, it presumes ED has power to prohibit,' the Bench said. The court had earlier asked the ED to produce material that justified the raids. When the agency submitted its evidence in a sealed cover, the Bench pointed out that the documents did not align with the submissions made by the ED. 'There's a divorce between the note and the submission,' Justice Ramesh observed. The ED later agreed to withdraw the notices and return the seized materials, including phones, laptops, and hard drives taken from Bhaskaran's Alwarpet residence. Orders on interim applications have been reserved, while the main petitions have been adjourned by four weeks. Bhaskaran, a relatively new name in the Tamil film industry, is the producer behind Dawn Pictures and is currently helming three of Tamil cinema's most anticipated films – Idli Kadai starring Dhanush, Parashakti featuring Sivakarthikeyan, and Silambarasan TR's STR 49. Collectively, these projects are estimated to be worth between Rs 400 and Rs 550 crore. His name came up on the ED's radar not just for the simultaneous high-value productions but also for the alleged proximity to DMK and Udhayanidhi Stalin. Of particular interest is Parashakti. Directed by Sudha Kongara and produced with a reported Rs 200 crore budget, the film is said to be a powerful political period drama centred on the anti-Hindi agitation in Tamil Nadu, with Sivakarthikeyan and Ravi Mohan in pivotal roles. The film was expected to release before the 2026 assembly polls. Its timing, amid national discussions on federalism, linguistic identity, and central overreach, has only sharpened speculation that Bhaskaran's cinematic ventures may also be seen as political expressions. The ED's operations came just months before Parashakti is expected to enter its final production phase. In their petitions, Bhaskaran and Ravindran claimed that neither had been named as an accused in any ongoing money laundering case. Senior counsel for Bhaskaran argued that the producer had no links to the TASMAC scam and that the seizure of his personal digital equipment was illegal. Ravindran, a director at Dawn Pictures, filed two petitions challenging the ED's sealing of his office at Semmenchery and a rented apartment at Poes Garden, arguing that the agency lacked legal backing for such action. The court, while granting the ED time to clarify its position, questioned whether notices barring entry into premises amounted to sealing. 'No sane person would ignore such a notice and walk in. That fear itself shows it is a de facto sealing,' said Justice Lakshminarayanan. The ED's counsel insisted that the search was based on 'credible information' and reiterated that Bhaskaran and Ravindran were not accused, but had failed to cooperate despite repeated attempts by the agency to contact them.
India.com
27-05-2025
- Entertainment
- India.com
5 Top South Indian Actresses Who Got Married Twice, Thrice: This 46-Year-Old Beauty Made Heads Turn With Her Sindoor Look At Cannes 2025
3 / 6 Lakshmi's first marriage was with Bhaskaran with whom she has a daughter, actress Aishwariyaa Bhaskaran. They later got divorced. Her second marriage was with co-star Mohan Sharma on the sets of Chattakari (1975) and ended in divorce (1980). Reportedly, while she was shooting En Uyir Kannamma (1988) she and actor-director M Sivachandran fell in love and got married (1987). The couple adopted a girl named Samyuktha in 2000.
