Open to opportunities

Sahab tariq

@sahabtariq

Message

I am a senior Machine Learning Engineer specializing in NLP, Computer Vision, and production ML systems.

Pakistan

Message

What I'm looking for

I seek roles where I can build and productionize NLP and computer vision systems, work on MLOps practices, and solve impactful, real-world problems within collaborative engineering teams.

I am a Machine Learning Engineer with over seven years' experience building and productionizing NLP and Computer Vision solutions. I have designed and deployed models for resume parsing and question answering, phishing detection with three production models, facial expression and emotion analysis for interviews, and real-time computer vision systems that improved performance from 5 fps to 33 fps.

I focus on end-to-end ML systems: data selection and representation, feature engineering, model training and tuning, and MLOps for reliable production delivery. I am comfortable researching new techniques, extending ML libraries, and building scalable services using tools such as TensorFlow, PyTorch, FastAPI, Docker, and Redis to solve practical problems in industry.

Experience

Work history, roles, and key accomplishments

Current

Sr. Machine Learning Engineer

Current

AppsGenii Technologies

Jul 2021 - Present (4 years 1 month)

Design and deploy ML and deep learning systems for resume parsing and Q&A on unstructured data; analyzed facial expressions, voice, and stress during interviews to enhance candidate assessment.

Python NLP Computer Vision TensorFlow PyTorch fastAPI MLOps OpenCV Data Analysis

Machine Learning Engineer

SlashNext

Jul 2018 - Jun 2021 (2 years 11 months)

Developed NLP models to detect phishing using TF-IDF and Random Forest and combined computer vision features with TF-IDF; maintained three models in production and reported precision and recall.

Python NLP TF IDF scikit learn Random Forest PyTorch Data Visualization

Computer Vision Developer

SIXLogics

Jun 2016 - Jun 2018 (2 years)

Built object detection and tracking systems, mapped world coordinates to 2D using homography, and trained InceptionV2/Keras classifiers, improving real-time performance from 5 fps to 33 fps.