Summary AI Professional with around 3.5 Years of functional expertise in preparing data, developing and deploying highly scalable machine learning models .Proficient in computer vision, deep learning, information retrieval, OCR and machine learning as well as scripting language like python.Proficient in application development and deployment using Python, SQL, Docker, Spring Boot, Jenkins, Kubernetes and AWS.
Senior Machine Learning Engineer
- Build shrinkage detection to identify theft. Used Yolo v5 for product detection and SORT algorithm for tracking the products and getting total count in entire transaction.
- Developed a deep learning based image classifier system that extracted information using object detection and image filters to draw inference on the validity of image.Enhancement of framework to eciently capture the capabilities of the processor for Computer Vision
- Developed object detection and multiple hypothesis tracking based algorithm for tracking of people in a video feed from a supermarket to data of checkout times for customers.
- Design and development of Computer Vision features for secure use cases such as face detection.
- Developed Python libraries for in-house MLOps platform’s Python SDK . Leveraged Docker, Kubernetes, CI/CD tools, MLflow, Kubeflow, and Kserve.
- Develop action plans to mitigate risks in decision making while increasing profitability by leveraging data science
- Utilized analytical and technical expertise in computer vision and machine learning to give insights on grading features
- Create custom APIs to fetch Forecasted KPI data with Confidence band and the anomalies in the forecasted data.
- Worked on various python scripts to obtain organized data using various pre-trained models based on various benchmarked dataset.
- Setting up pipeline using various AWS services for real time model predictions (Kinesis, Lambda, S3, Athena, QuickSight)
Lead ML Engineer Ecommerce Company
- Data processing of continuous/ discrete data, data cleaning, identifying appropriate models for the data.
- Deployment of a Deep Learning model for denoising distance images delivering twice better performance than the existing algorithm.
- Developed advanced statistical and AI / Machine Learning models forhigh- dimensional data using algorithms like Neural Networks, Linear/Polynomial/Logistic Regression, Multi-class classification, SVM, KNN, PCA, andCollaborative filterin
- Trained a model in tensorflow which can classify 60 different generic locations using state of the art algorithm Inception-V3 resulting in the accuracy of 90.5%.
- Build a scalable recommendation system to help sales representatives find the right products for the customers based on orders & digital footprint data
- Extracting insights like sentiment, attributes, opinions from videos using feature engineering and ML models.
- Defined and collected data, applied augmentation and established baseline to ensure clean and consistent data
Data Science Engineer Technology Company
- Training Unsupervised deep learning (Deep clustering: network with clustering) with own data for replacing commonly used pre-training network.
- Developed an automated anomaly detection framework using a statistical approach of shifting average as a short-term solution. Built unsupervised Clustering and Isolation Forest ML model to be a robust and long-term solution achieving a precision of 70%.
- Experience in building and fine-tuning OCR pipelines. Developed AI engine takes files in any format and extracts required key value pair
- Optimized legacy codes and various data-pulling tasks or downloading large files from a database/server using multiprocessing and Numpy vectorization, reducing the runtime by 90%.
- OpenCV based classification of receipts paid by customer and verification with code reader data.
- Data pipeline to extract product details, availability, price & promotions information daily from internal SOAP APIs supporting client’s website for various regions and time zones
- Analyze structured and unstructured data at scale to derive new insights and opportunities
- Pattern recognition and skew correction using OpenCV python.
- Contribute to internal research and development efforts in cutting edge areas including NLP, graph analytics, and deep learning / AI.
- Architected and implemented ML training & evaluation infrastructure in Python using OpenCV, scikit-learn, TensorFlow, and Keras
Skills
Language : Python (Flask, Django), Java, J2EE (Spring, Hibernate), Bash/Shell scripting, R.
Machine Learning: Supervised & Unsupervised Learning, Regression techniques, Linear Classifiers, K-NN, SVM, Tree based classifiers, Bagging & Boosting techniques, Regularization, Clustering, Dimensionality Reduction techniques, HMM, CRF
Deep Learning: DNN, CNN, RNN, LSTM, Word Embeddings, Categorical Embeddings, Encoder- Decoder, Anomaly Detection, Attention based architectures and Transformer Based Architectures
Text Analytics: Aspect Based Sentiment Analysis, Topic- Modelling, Machine Translation, Text Classification, Text preprocessing, Regular Expressions
Platforms: Cloud (Azure Cloud, Google Cloud, AWS), Windows, Ubuntu, RHEL
Natural Language Processing :Text Clustering and Classification – word2vec, BERT, Alexnet. Text summarization
Model: Churn model , Spend prediction , Credit Risk model , Health Prediction , Product Recommendation model , Oer personalisation model , Acquisition model , Best Product model , Merchant named entity recognition , Preferred payment analysis , Fraud detection
