Maqbool portrait

Hi, I am Maqbool

In love with design and A.I

Heavily into on-device LLMs

Building an startup atMerantix, Berlin

Experience

Data (Knowledge) Engineer

May 2022 - Jan 2026

Taking part in laying a foundation of a data platform using Knowledge Graphs that will enable the digitalization of the construction industry using machine learning. So Python, Notebooks, RDF, SPARQL & Ontologies are my day-to-day business.

RDF-to-Text generation — (Grade 1.5)

February 2021 - September 2021

Researching Natural Language Generation problem of graphs (RDF) to text using state-of-the-art machine learning methods. Developing my solution in pipeline fashion using Pytorch and WebNLG corpus. Under the supervision of Prof. Dr. Axel-Cyrille and Dr. Diego Mousallem.

Student research assistant

July 2020 - September 2021

Developed open source chat bot for data scientists IDA. My contribution largely involved setting up project using Docker, creating chat bot functionalities using Java and at the end of the project used Python's Pandas and Flask for dataset upload and preprocessing functionality.

Software engineer

2018 June - 2020 June

It was my first working experience abroad (Germany) and in a startup. Here I contributed in their web app and worked with Angular, Typescript and Firebase as our real time database.

Software engineer (Cloud Security)

2016 November - 2018 March

Worked their mostly as a front-end developer using Angular 1 but was also involved in some Python and NoSQL tasks. Overall learned a lot about software development processes and best practices.

Nodejs developer cells

2019 February - 2019 October

Worked with NodeJS to develop rest APIs and migration scripts.

Selected Projects

An VLLM-powered solution to automatically perform actions on the computer based on the user query. Read more here.

Github

Grade 1.3

A chat-bot for data-scientists. Co-lead the team of 8 fellow students and also developed visualization and Natural Language Engine.

Github

A project about Neural Machine Translation or Sequence-to-sequence models. All models are written from scratch in order to be easily understood and integrateable in any future work!

Github

In this project, I am implementing basic tasks of NLP using statistical, ML and Deep learning techniques!.

Github

Tried to solve the problem of grouping results of an ambiguous query and how can we advance search engines! Created full fledged web crawler using Python from scratch. In this project, we used ElasticDB to index and store crawled web pages.

A naive fact checker! which extracts subject, object & predicate from the given sentence and try to validate fact using DBpedia knowledge base.

Github

The Simplest Java Code Editor, coded in Java.

Website

A Rapid MS Word Document changer.

Website

Even more about me ..