IMPRS-CS Scholar

• Oct 2020 - Present

Fully funded 3 semesters of graduate school by the International Max Planck Research School for Computer Science (IMPRS-CS) which is a graduate program jointly run by the Max Planck Institute for Informatics (MPI-INF), the Max Planck Institute for Software Systems (MPI-SWS), and the Computer Science Department at Saarland University. The IMPRS for Computer Science (IMPRS-CS) is highly research-oriented. Fully funded 3 semesters of graduate school by the International Max Planck Research School for Computer Science (IMPRS-CS) which is a graduate program jointly run by the Max Planck Institute for Informatics (MPI-INF), the Max Planck Institute for Software Systems (MPI-SWS), and the Computer Science Department at Saarland University. The IMPRS for Computer Science (IMPRS-CS) is highly research-oriented.

Graduate Teaching Assistant

• Oct 2021 - Present

Tutor for NNTI: Neural Networks Theory and Implementation Tutor for NNTI: Neural Networks Theory and Implementation

Graduate Research Assistant

• Sep 2021 - Jan 2022

Working on a haptic toolkit for VR. Working on a haptic toolkit for VR.

Data Scientist (NLP)

• Apr 2021 - Aug 2021

Worked on:1. Huggingface's first pre-trained domain-specific Chemical Industry language model (BERT) using masked language modeling, by first creating a domain-specific text dump of 9+ million passages from 50k+ long multipage documents and Wikipedia. (https://huggingface.co/recobo/chemical-bert-uncased)2. Finetuned the domain-specific language model on extractive question answering task.3. Structured information extraction from heterogeneous multi-lingual data sources, used OCR and text mining techniques to assemble pipelines.4. Worked on and delivered Tabular Question Answering on data at scale feature since current SOTA approaches like TAPAS don't scale to big datasets.5. Introduced knowledge graphs (Neo4j) based feature to enhance the search for queries involving one or more Named Entities, used Machine Learning for entity relationship identification.6. Participated and presented in demos of Recobo search engine with both community-based sessions and client PoC demos.7. Trained a Transformer based classifier for the Pharmaceutical/Chemical industry domain.

Data Scientist

• Oct 2019 - Oct 2020

The role concretely required me, in the IT Infrastructure domain for IBM's Multicloud Management Platform(MCMP):1. Read Machine Learning papers & spin up experiments and prototypes2. The use cases cover pervasive ML techniques for solving NLP & Anomaly detection problems.3. Develop baselines and compare them against state-of-the-art techniques.4. Patented one of our research projects as a co-inventor. The role concretely required me, in the IT Infrastructure domain for IBM's Multicloud Management Platform(MCMP):1. Read Machine Learning papers & spin up experiments and prototypes2. The use cases cover pervasive ML techniques for solving NLP & Anomaly detection problems.3. Develop baselines and compare them against state-of-the-art techniques.4. Patented one of our research projects as a co-inventor.

Big Data Analytics Consultant

• Sep 2016 - Aug 2019

Worked with North and South American Multinational clients namely Citi Bank, Wells Fargo, Delta Airlines, Pfizer, Banorte Bank, DOW Chemicals, Kaiser Permanente, Bank of America to name a few, whilst participating in R&D and deployment of the following use cases:1. Design of architecture and implementation of serverless self-service data analytics platform based on ELK stack on IBM cloud. Tools: Dockers, Kibana, Elasticsearch, Logstash, Nginx2. Design architecture and implement a big data text analytics platform based on IBM Watson Explorer on-premises for service desk data. Tools: Ruby, Javascript, SQL, Linux, Apache Lucene, IBM Watson Explorer3. Developed a fully customized open-source big data search analytics solution based on Elasticsearch and designed custom plugins using Highcharts API, moreover integrated with IBM OIDC identity authentication for big data sets of service desk data. Tools: Elasticsearch, Kibana, Logstash, Highcharts, Node JS, Angular JS, Linux, IBM OIDC, Nginx4. Automated data pipelines based on Logstash for data refresh on Elasticsearch with minimum down-time. Tools: Logstash, Bash5. Predictive Models development in IBM SPSS and Python( Use cases like SLA classification, Ticket resolution prediction, etc. for service desk data) Tools: Python Flask, Scikit learn, Numpy, Pandas, Scipy, IBM SPSS6. Real-time Social Media Analysis which included sentiment analysis, entity, and concept extraction from social feeds of client's social accounts. Tools: Java, IBM Watson7. PoC of Serverless Chatbot for Etisalat which integrated with their mobile app. Tools: Node Js.8. PoC with PITB on flood forecasting dashboard, which forecasted water levels at different barrages using a statistical model. Tools: IBM Weather API, Python, IBM Watson Analytics

Internship

• Jul 2016 - Aug 2016

Java Developer

• Jun 2016 - Jun 2016