About me

I’m working as datascientist past 2 years, I like to solve challenging complex problems and provide optimal solution to grow the business.

Following projects are i worked on:

1. Churn prediction and performance attrition - Classification and Regression
   • Used Survival analysis, HMM, Tree based models are Randomforest, XGBoost and RNN
2. Interpreting models - Explaining prediction on individual records
   • Shapley values, Lime, Eli5
3. Chat Bot - IT Help desk, Health Bot
   • Pipeline of Tokenization, POS tagging, Stop words removal, Stemming and lemmatization
   • Intent classification
   • Entity Indentification
   • Build a minimal Utterence generator
4. Image classification and Object recognition
   • Used Inception v3 model and retrained last layer of softmax
   • YOLO
5. Bill details extraction through invoice using machine learning (No Regex)
   • Tesseract to get content and bounding box coordinates
   • Applied KNN on coordinates to get related content
6. Anomaly detection
   • Used HBOS, Denoising Autoencoder, Decision Tree
   • Build Api Server using spray and worker node by Undertow
7. Handwritten tamil character recognition
   • Used opencv to extract content and region of bounding box
   • Using Knearest found the closely matched letter

Technolgy Stack:

Languages : Python, Java, C

DataBase: SQL, MongoDB, HBase

Familiar with: ELK stack, Redis, Spark

Libraries: Numpy, Tensorflow, Scikit-learn, Spark-ml, Pandas, Open-CV

Devops: VSTS, Git, Jenkins

Cloud: Azure, GCP

Apart from that datascience, i worked on following during my college

1. Developed weather app in android
2. Teacher tool - Managing student attendance and course materials
3. Crime Control app - social network for sharing crimes happening around us and informing others. And Geotagging the crimes.
4. Prestashop extension - Open source Ecommerece platform

On Security wise,

1. Found a security vulnerability on ISP, reported back to them.
2. Reported sensitive information leakage and security threats within organisation.