M Vishnumanikandan

I specialize in machine learning, data analysis, and backend development. With one year of experience at AiSpry, I have optimized data pipelines, built predictive analytics models, and delivered scalable AI-driven backend solutions. I was deployed onsite at the National Centre for Statistics and Information (NCSI), Oman Tax Authority, collaborating with government stakeholders on socio-economic prediction systems powered by geospatial intelligence.

Associate Data Scientist - Ai Spry - Chennai

(2024-12 - 2025-11)

Client: North Western Railways

• Built an end-to-end audio intelligence system leveraging Google Gemini 2.5 Pro to transcribe Hindi railway announcements, identify operational keywords, and generate timestamps + frequency metrics for real-time insight.
• Designed an automated audio preprocessing and transcription pipeline powered by FFmpeg (PCM mono conversion), SpeexDSP (audio enhancement), and Pydub (chunking) orchestrated via cron jobs for continuous processing.
• Integrated AWS S3 for scalable audio storage and PostgreSQL (AWS RDS) for structured metadata and analytics, enabling fast search, tracking, and audit of announcements.
• Skills: Python, Flask, FastAPI, React, JavaScript, AWS S3, AWS EC2, PostgreSQL, Google Gemini, Speech-to-Text, Keyword Spotting, Audio Processing (SpeexDSP)

Associate Data Scientist - Ai Spry - Chennai

(2024-12 - 2025-11)

Client: NCSI Oman (National Centre for Statistics and Information)

• Performed on-site Exploratory Data Analysis (EDA) on Oman Tax Authority datasets, generating automated profiling reports using Sweetviz, computing correlations and mutual information to uncover key drivers of demographic consumption patterns.
• Processed VIIRS night-time satellite imagery in QGIS, using Oman shapefiles, raster clipping, and masked extraction (H23V06 tiles) to isolate the country's night-light intensity zones.
• Computed zonal statistics to quantify illumination levels across administrative regions, enabling predictive modelling of household income and expenditure using night-light intensity as a socio-economic proxy.
• Skills: Python, QGIS, VIIRS Night-Time Data, Geospatial Analytics, EDA, Machine Learning

Backend Developer - Saturn Pyro - Chennai

(2024-07 - 2024-08)

• Developed modules for MyESG Online, a web-based platform enabling organizations to record and manage Environmental, Social, and Governance (ESG) metrics and compliance data.
• Implemented user authentication, role-based access, and ESG data submission workflows using Django, ensuring secure onboarding and structured data capture.
• Integrated AWS S3 for reliable storage of ESG documents and evidence files, enabling scalable.

Data Science Intern - Excel R - Bangalore

(2024-02 - 2024-06)

B.Tech. - Artificial Intelligence and Data Science - Anna University, KCG College of Technology (2024-05)