Sami Ullah

About

Strong background in Data Science (Machine Learning, Deep Learning and Natural Language Processing). Experienced in implementing a wide range of algorithms (both traditional machine learning models and deep learning architectures) in academia and industry. Technical skills: - Data Science - Python(Pytorch, Tensorflow, Sci-kit Learn, Pandas, Numpy etc), R ,and Java - Machine Learning - Deep Learning - Natural Language Processing - Bayesian Deep Learning (reasonable understanding). - IBM Watson Studio - Microsoft Azure ML - Statistical analysis - SQL

Experience

• Data Scientist · IBM05-31-2018 – 05-28-2022

  - Manager's Choice Award 2018 - Expertise in Advanced Analytics (Python, R) with an understanding of data mining and business problem definition. - Main duties and responsibilities for implementing ML use cases in bank(Allied Bank Limited)(An IBM client): 1. Meet customer to understand business needs 2. Design specific data science solutions to solve selected problems 3. Code and test the proposed solution (PoC) 4. Communicate results to stakeholders 5. Provide final reports of the work done and related results 6. Use Cases: (Allied Bank has a customer base of ~4 million) ---- Customer Behavioral Segmentation, ---- Next Best Action, ---- Churn Prediction,Cash Optimization, ---- Real-Time Fraud Detection - Other tasks at IBM for State of Ohio and State Farm Projects: 1. Extensive R&D on state of the art Natural Language classifiers such as FastText, Watson NLC, Sequence to Sequence models, LSTM, ELMO, fast.ai. 2. Built a domain specific chatbot by training a seq2seq model and Rasa NLU for customer issues resolution. 3. Trained BiLSTM-CRF NER, BERT, utterance clustering, text classification(CNN/LSTM). 4. Developed scripts for cleansing, processing and extracting only useful information from unstructured text using NLP techniques. 5. Developed Restful APIs for facilitating different use-cases in production. 6. Predictive Analysis of Time Series data with auto-regression techniques and stacked LSTM's. 7. Knowledge Graph- Taking the research of Google and papers like NELL, YAGO, and DeepDive to a next level to cater complex sentences, with complex form of entities. Enhancing the concepts of Knowledge Graphs to fit for any sort of data and extract information without any fixed type of relationships or entities by making use of the state of the art techniques like Syntaxnet. 8. Build our own word2vec model to check the similarities and injest them in knowledge graph.

Education

• University of San FranciscoBachelor's Degree, Computer Science

Similar talent on Pangea

Srijit D.Lead Data Scientist | Machine Learning ExpertBengaluru, KarnatakaSolon D.AVP - Data Engineering @SwissRe | Palantir Foundry | Databricks | Azure | Blogger | MentorBengaluru, KarnatakaGGoutam J.Data Scientist / Manager, ML & Data Pipelines | 5 Years Experience (AWS/Azure, NLP, CV)Kolkata, West BengalAbhishek K.Expert Data Engineer & Analyst Specializing in AI-driven Solutions and Business IntelligenceBengaluru, Karnataka