Responsible for the Adverse Media product and I supervised a team of five junior data scientists, solving numerous NLP challenges with Transformers, LLMs and advanced NLP algorithms driving business revenue growth.Deployment & Scalability: Managed the entire ML model lifecycle, from development to deployment. Scaled production APIs to handle millions of requests, ensuring high performance.Technical Proficiency: Utilized Docker, Kubernetes, Elasticsearch, FastAPI, Apache Airflow, and Neo4j to optimize workflows and project outcomes.Extra Responsibilities:Implemented SEO strategies for an AML check project, resulting in Top rankings for most target keywords and a 200% increase in organic leads

Worked on Adverse Media Module:Created the parser which monitors 5000 + news sources 24/7 to collect present and past data.Solved NLP problems related to unstructured data and built models for: Entity Segmentation,Entity Recognition,Entity-based Sentiment Analysis,Relationship Extraction,Information ExtractionUtilized BERT, RoBERTa, DeBERTa, SpaCy, and Stanford NLP.Stored collected data in a structured format using MongoDB.Worked with Elasticsearch, FastAPI, Logstash, Kibana, Docker, and Kubernetes to serve this data to clients.Extra Responsibilities:Managed the annual event for the organization, making it interactive and collaborative.

Completed Mentorship program as mentor to provide guidance and grow up the skills in Python Programming Language of students that are enrolled in the course provided by @TopTroveFoundation

Successfully completed the client’s project titled “E-commerce store review text classification using Deep Learning techniques with Cloud Deployment”.Converted data into actionable insights by predicting and modeling future outcomes.Solved various ML/DL/NLP problem statements provided by Forsk Technologies.

Worked on Amazon Rekognition/Python/Deep Learning/OpenCV to make a real time project "Face Mask Detection" .This project uses a Deep Neural Network, more specifically a Convolutional Neural Network, to differentiate between images of people with and without masks. The CNN manages to get an accuracy of 98.2% on the training set and 97.3% on the test set. Then the stored weights of this CNN are used to classify as mask or no mask, in real time, using OpenCV. With the webcam capturing the video, the frames are preprocessed and and fed to the model to accomplish this task. The model works efficiently with no apparent lag time between wearing/removing mask and display of prediction.The model is capable of predicting multiple faces with or without masks at the same time

Accumulated data that can be helpful for marketing team and our public relation team and get connected with many new people who can provide profit to SkillsRediscovery in terms of business and relationWorked closely with two initiatives “Talk with Inspiration” and “Coffee with Coach” .In “Talk with Inspiration” we connected various YouTubers with SkillsRediscovery while in “Coffee with Coach” we connected various Coaches and Mentors with SkillsRediscovery