Open to opportunities

mavericks prj

@mavericksprj

Message

I build production machine learning systems for NLP, recommendations, and computer vision.

Pakistan

Message

What I'm looking for

I'm looking to build and deploy impactful production ML systems, especially across NLP, recommendation engines, computer vision, and MLOps, while working with a collaborative engineering team.

I've built and deployed production ML systems at NovaTech Analytics, including a transformer text-classification pipeline processing over 2 million documents monthly and improving categorization accuracy from 81% to 94%.

I also designed a real-time recommendation engine that increased click-through rate by 18%, reduced training costs by 35% through spot GPU migration, and set up MLflow experiment tracking and a model registry for a six-person ML team.

Across BrightWave Labs and PixelForge Robotics, I've fine-tuned BERT for support-ticket sentiment analysis, automated high-volume Airflow pipelines, deployed containerized inference services on Kubernetes, and developed a YOLOv5 defect-detection model reaching 92% precision.

Experience

Work history, roles, and key accomplishments

Current

Machine Learning Engineer

Current

NovaTech Analytics

Jan 2024 - Present (2 years 7 months)

Built and deployed a transformer-based text classification pipeline processing 2M+ documents/month, improving accuracy from 81% to 94%. Designed a real-time recommendation engine increasing click-through rate by 18% and reduced training costs by 35%.

PyTorch TensorFlow Transformers MLFlow Docker Kubernetes AWS Sagemaker Python SQL Spark Airflow CI CD

Machine Learning Intern

BrightWave Labs

Jun 2023 - Dec 2023 (6 months)

Fine-tuned a BERT-based sentiment analysis model on customer support tickets, reducing manual triage time by 40%. Built Airflow data pipelines for automated ingestion and preprocessing of 500K+ daily records.

BERT Hugging Face Transformers Airflow Docker Kubernetes Python SQL

Data Science Intern

PixelForge Robotics

Jun 2022 - Aug 2022 (2 months)

Developed a YOLOv5 computer vision model for defect detection on a manufacturing line, achieving 92% precision. Conducted statistical A/B tests to evaluate model rollout impact and presented findings to stakeholders.

YOLO V5 Computer Vision A B Testing Python Statistics

Education

Degrees, certifications, and relevant coursework

COMSATS University Islamabad, Lahore Campus

Bachelor of Science, Computer Engineering

Pursued a Bachelor of Science in Computer Engineering with coursework in Data Structures & Algorithms, Database Systems, Signals & Systems, Machine Learning, and Linear Algebra.