Mohammed Junaid Anwar Qader

šŸ‘‹ Hello! I'm Junaid, and I am currently pursuing my MScAC at the University of Toronto with a concentration in Artificial Intelligence. I completed my undergrad at NIT Warangal, where I published research on acoustic mosquito event detection using deep learning and worked on anti-cathepsin prediction for drug discovery as my final-year project.

Iā€™m passionate about end-to-end AI systems, from research to real-world deployment. My current focus is on Implementing and fine tuning large language models (LLMs) and building retrieval-augmented generation (RAG) pipelines.

šŸ” Current Interest: Iā€™m deeply exploring the LLM + RAG space, designing pipelines that combine semantic chunking, vector search, and conversational memory to enable domain-specific Q&A. Alongside this, Iā€™m learning the nuances of LLM training and fine-tuning, and applying these skills to build systems that bridge the gap between raw research and production-ready applications.

šŸŒ Career Aspiration: I aim to contribute to projects where AI makes tangible impact, whether in drug discovery, healthcare, or enterprise AI systems. I thrive on learning by building, diving into papers, and translating cutting-edge research into meaningful applications.

šŸ”— Letā€™s Connect: If youā€™re working on something exciting with LLMs, RAG, or deep learning in healthcare, Iā€™d love to connect. Always open to collaborations, knowledge-sharing, and opportunities to build impactful AI solutions.

Education

University of Toronto

Master of Science in Applied Computing (MScAC)
AI Concentration

Relevant Coursework: Computational Imaging, Neural Networks and Deep Learning, AI for Drug Discovery, Visual and mobile computing systems.

September 2025 - December 2026 (Expected)

National Institute of Technology, Warangal

Bachelor of Technology
Computer Science and Engineering

GPA: 9/10

Relevant Coursework: Computer Networks, Machine Learning, Artificial Intelligence, Operating Systems, Database Management systems, Data Structure and Algorithms, Advanced Algorithms, Software Engineering, Theory of Computation, Object Oriented Programming, Software Testing, Integral Calculus and Transforms, Probability Statistics and Queuing Theory, Computer Architecture, Mobile Computing

August 2021 - May 2025

Industry Experience

Wells Fargo

SDE Intern
  • Designed and developed a chatbot plugin for VS Code and IntelliJ IDEA, leveraging NLP techniques (RAG model, cosine similarity, and vector databases) to solve developer queries.
  • Created an extension for Harness API to display all development logs and pipeline states, integrated seamlessly into VS Code and IntelliJ IDEA, reducing context switching for developers, boosting productivity by 40%.
  • Integrated Confluence API with the chatbot extension, enabling real-time, NLP-driven query resolution and quick access to relevant documentation.
  • Employed JavaScript, VS Code API, Java AWT Swing, NLTK, SpaCy, and GitHub Actions API to enhance the chatbot and Harness API extensionā€™s functionality.
  • Collaborated with cross-functional teams to ensure security compliance, meeting Wells Fargoā€™s protocols while optimizing the chatbot and extensionā€™s usability for developer productivity.
May 2024 - July 2024

Research Experience

Molecular Descriptors in Drug Design

Research under Prof. Tapan Kumar S.
  • Research project as part of the final-year B.Tech curriculum, focusing on predicting the activity and release patterns of chemical molecules in the human body using molecular descriptors.
  • Applied feature selection techniques like Recursive Feature Elimination (RFE), Forward/Backward Selection, and Gradient Boosting to optimize descriptor selection and improve model transparency.
  • Explored anti-cathepsin activity prediction using molecular structures, leveraging feature selection algorithms to identify key descriptors from high-dimensional datasets
  • Applied feature elimination techniques to obtain an optimal descriptor set and trained 1D CNN models with SMOTE to address class imbalance, achieving 97% accuracy in identifying potent inhibitors.
Aug 2024 - Apr 2025

Deep Learning and IoT-based Mosquito Wingbeat Event Detection

Research under Prof. Venkanna U.
  • Collected raw mosquito and environmental sound data across diverse locations, creating a comprehensive dataset for model training.
  • Created an artificial dataset by combining environmental noise and lab-recorded mosquito sounds, addressing data scarcity challenges. Applied data augmentation techniques such as time stretching, time shifting, random gain addition, and mix-upto generate diverse, environment-specific training data.
  • Preprocessed audio data using log-mel spectrograms and Per-Channel Energy Normalization (PCEN), enhancing model robustness in noisy environments.
  • Developed and trained a custom CNN-Temporal Convolutional Network (TCN) architecture, achieving over 90% accuracy in detecting mosquito wingbeats.
  • Implemented attention mechanisms to improve model focus on critical acoustic features, optimizing mosquito detection across varying environmental conditions.
Jan 2024 - May 2024

PUBLICATIONS

  1. M. J. A. Qader, C. M. Sah, T. K. Sahoo, S. K. Majhi and K. Mishra, "CathepsinDL: Deep Learning-Driven Model for Cathepsin Inhibitor Screening and Drug Target Identification," in IEEE Access, doi: 10.1109/ACCESS.2025.3617246. https://ieeexplore.ieee.org/document/11192524

  2. Seervi A, Mulani N, Anwar MJ, Venkanna U. "IoT-Enabled Intelligent Framework for Real-Time Mosquito Detection and Monitoring." SN Computer Science. 2025 May 21;6(5):484.. https://link.springer.com/article/10.1007/s42979-025-04016-y

Skills

AI/ML
  • Classical ML/DL,
  • Computer Vision,
  • LLMs,
  • Agentic Systems,
  • Natural Language Processing.
Programming Languages
  • Python,
  • Java,
  • C/C++,
  • Javascript.
Libraries/Frameworks
  • NumPy,
  • Pandas,
  • Matplotlib,
  • Sklearn,
  • PyTorch,
  • Flask,
  • Django,
  • React.
Database
  • MySql,
  • ChromaDB (vector)
MLOPs
  • Git,
  • MLflow,
  • AWS,
  • SageMaker,

Projects

This is a collective dump of all my projects, big and small. Some of them were part of coursework, others for self learning.

  • MLOps Weather Predictor
    • End-to-end weather prediction pipeline for Toronto, forecasting weather codes using historical time-series data.
      • Implemented a full MLOps pipeline with automated retraining, deployment, and monitoring to maintain production-ready performance. Engineered time-series features from historical weather data and trained an XGBoost model to predict weather codes.
  • SLM-NanoGPT
    • A lightweight implementation of the NanoGPT architecture for text generation tasks.
      • Implemented a small-scale LLM from scratch and trained it locally on the TinyStories dataset, gaining hands-on experience with tokenization, dataset preparation, and pretraining. Currently focused on learning more about advanced LLM architectures and techniques.
  • CaptionCrafter
    • Generates catchy captions for Instagram posts
      • Developed a deep learning pipeline using DenseNet201 (CNN) for feature extraction and LSTM for sequential caption generation, trained on the Flickr8k dataset to generate captions for images. Integrated with Groq Cloud API to generate dynamic hashtags and captions optimized for social media platforms
  • SmartApply
    • AI-powered job application assistant
      • Developed a pipeline to scrape job portal data and process user-uploaded resumes, leveraging NLP techniques for context-aware data extraction and analysis. Engineered prompts to identify recommended skills and suggest tailored project ideas to enhance the userā€™s application. Provided constructive feedback on resumes and a job summary feature, helping users gain actionable insights for improving their applications.
  • Selfie-Anime
    • Generates anime-style portraits from selfies using GANs. A part of CS315 Artificial Intelligence course.
      • This project implements an image translation model using CycleGAN to convert selfie images into anime-style representations. The CycleGAN framework is trained to learn mappings between two distinct image domains without requiring paired datasets.

Awards & Honors

  • NIT Warangal Institute Merit cum Means Scholarship Recepient of the scholarship for the academic year (2023-2024) for being in the top 10/160 in the department.
  • NIT Warangal Institute Merit cum Means Scholarship Recepient of the scholarship for the academic year (2022-2023) for being in the top 10/160 in the department.

Volunteering

  • Member Technical AI/ML Faction

    Software Developer Club, NIT Warangal

    Created AI/ML problem statements for the SDC hackathon during Technozion, contributing to the eventā€™s success, mentored participants at Technozion, providing guidance on problem statements and solutions.

    August 2024 - April 2025