Oh! I’ll be back, Let my work talk to you.
✳️
Meet Rohit Panicker
Meet Rohit Panicker
Featured Project
FlowLogs Intelligence: AWS VPC Monitoring A JAVA agent that would periodically retrieve VPC Flow Logs stored in S3 buckets and parse the information to keep track of IP traffic to and from network interfaces inside VPC.
Payload Compression: G-ZIP payload compressions are implemented for transmitting data to any custom APIs having payload size.
WatchDog-Service: Service and configuration files are continuously monitored to detect changes in environmental or configuration
parameters, instantly refreshing the application context and avoiding the need to restart the application
restrictions.
K8s/Docker Version: Built as a Multi-Cloud containerized version, deployed with the help of helm charts and could be used for AIOps,
Network-Flow Visualization or for Analyzing Network Interfaces in VPC.
Logs Intelligence
-> AiOps, Traffic Visualization, Threat Detection
Multi-Cloud containerized version
-> Deployed using Helm Charts
Adapter Design
-> Dynamic Adaptation : AWS, GCP, Azure
Additional Projects
Natural Language Processing
Summarized Speech Notes
Operational and Cross-Platform Audio Processing Layer: To provide audio inputs and to allow users to move about the application, the application's GUI is constructed using the Tkinter python module. The pyaudio module has been used to facilitate cross-platform audio processing and recording operations, and frames of audio are constantly appended to produce a final .wav file for additional processing.
Speech-To-Text Processing Layer: Recorded audio is first cleaned of background noise using the SpeechRecognition python module and is later converted to text by utilizing Googles’ Speech Recognition API.
Summary Generation Layer: This layer implements Language processing methods such as tokenization, weighted frequency, sentence tokenization, sentence scoring and sentence selection for generating the final text based summarized output.
-> VIEW RESEARCH PAPER FOR SPEECH NOTES
Cross Platform Monitoring Intelligence
Interpreter : Microsoft Open Management Infrastructure
Operation: Devised and implemented a JAVA interpreter that could fetch query results from OMI’s system provider installed on one’s machine and convert unstructured results into a structured Java Collection (ArrayList or List of ArrayList).
Multilevel WMI/CIM instance parser: Parses single, multiple, and hierarchical CIM/WMI instances to facilitate remote monitoring of any windows device over WSMan protocol.
Result Formats: Results -> JSON, XML or other Data-Interchange formats.
Machine Learning
Strategical scheduling of tasks using Genetic Algorithm
Algorithmic Model: Designed a model using genetic algorithm, to create a timetable by scheduling tasks like arranging different subjects in an optimized manner by generating more than 10000 timetable samples and choosing the best fit result.
Hard-Constraints: There should not be any single instance of a faculty taking two classes
simultaneously, A class group must not have more than one lectures at the same time.
Soft-Constraints: More or less equal load is given to all faculties, Required time (hours per week) is given to every batch.
PUBLICATIONS
Audio Data Summarization System Using Natural Language Processing
International Research Journal of Engineering and Technology
This paper presents techniques for converting speech audio file to text file and text summarization on the text file. For the former case, we have used Python modules to convert the audio files to text format. For the latter case, Natural Language Processing’s modules are used for text summarization. A Python toolkit named SpaCy is used for the English data functions. Summarization method involves important sentence obtained when the extraction is investigated. Weights are assigned to words according to the number of occurrences of each word in the text file. This technique is used for producing summaries from the main audio file.
Like what you see? Hire me