Ashima Chawla

• Developed and maintained courteous and effective working relationships.
• Design AI-driven dashboard (jupyter based) solutions that are customer-focused, to translate data into useful information.
• Auto-tune model using Bayesian Hyperparameter optimisation technique.
• Working in collaboration with European Horizon MonB5G H2020 project.
• Working on Interpretable Anomaly Detection using Deep Learning (Tensorflow, Keras) to monitor network management. Deep learning algorithms include autoencoders, LSTM, seq-to-seq models, probabilistic based models.
• Conduct research and develop prototypes and proof of concepts to work on data sets of sparse high dimensional data; perform pre-processing and analyze data to gain actionable insights.
• Strong verbal and written communication skills with the ability to tell compelling stories to suit clients/stakeholders with technical knowledge.

• We use Deep Neural Networks (DNNs), a powerful tool for anomaly detection in cybersecurity (intrusion detections systems) and in telecommunications network management.
• We demonstrate the effectiveness of unsupervised neural network algorithms to detect univariate, multivariant anomalies and anomalous features.
• Design computationally efficient DNN based algorithms.
• To overcome the “Black Box” nature of neural networks (and thus increase their acceptance by domain experts), we investigate in detail to provide root cause analysis (interpretability) to explain the end user the output of a neural network model
• Experience with cloud-based infrastructure i.e. AWS to perform real-time anomaly detection for IoT devices.
• Conducted research, gathered information from multiple sources and presented results.

Team Size 7, 4.1 (Admin and development), Dashboard reporting

• To investigate Ericsson real network performance management datasets by providing insights using Analytical reporting tools
• Experience with the BI tools Tibco Spotfire Analyst, Web player and Business Objects
• Direct interaction with customers and developed LTE based network reports, RAN Performance overview, and Two-Way Active Measurement Protocol (TWAMP) report for customers
• Experience in creating Reports for ENIQ Statistics using Tibco Spotfire Professional which involves custom expressions, IPython scripts
• Customer Experience - Ericsson LMI ; January 2015 – August 2017 (Short Term Visits-3 months)