< HELLO WORLD />

Awais Tahseen

Results-driven AI Engineer with strong expertise in traditional machine learning, deep learning, hypothesis testing, and end-to-end ML solution development. Skilled in Python and experienced in building practical AI systems across Computer Vision, LLMs, and Generative AI frameworks such as LangChain, LangGraph, and CrewAI. I focus on understanding the real problem before writing code asking “why,” validating assumptions, and identifying whether AI is truly the right solution for the business challenge. With hands-on knowledge of MLOps tools and practices including Docker, Kubernetes, and Terraform, I aim to build AI solutions that are not only intelligent but also scalable, reliable, and production-ready.

< EXPERIENCE />

Career Break & Continued Professional Development

Founder

Sep 2025 – Present

Currently, I am not engaged in a full-time role and am using this period intentionally to spend quality time with my family, learn a new spoken language, further strengthen my technical skills, and work on an AI-focused product. I remain actively open to new opportunities where I can contribute, grow, and build impactful AI and software solutions.

Tkrupt, Lahore (Remote)

ML Engineer

Dec 2024 – May 2025

Developed and deployed ML models, chatbots, and AI solutions using Langchain, OpenAI, DeepSeek, Huggingface, vector databases, AI agents, and integrated low-code/no-code platforms like Zapier and n8n, with PostgreSQL and Airtable for data management.

NESCOM, PK

Machine Learning Intern

Sep 2024 – Nov 2024

Built Transformer-based models by using Pytorch and leveraged advanced techniques like Wave2Vec2.0 for nuanced audio feature extraction instead of mel-filter banks. Optimized classification with ResNet-1D, Transformer Encoders, and BiLSTM layers, while creating custom datasets with fake audio segments to improve model robustness and minimize error rates

TruID NSTP– NUST H12 Islamabad, PK

Machine Learning Intern

Jul 2024 – Sep 2024

Annotated images using CVAT.ai to create high-quality labeled datasets for training YOLO models to extract features like flags, chips, and holograms , Prepared and managed datasets while addressing challenges like data imbalance and also learnt about Perspective Warping and homography, Trained and optimized YOLO models to detect features of the card and trained classifier in tensorflow improving metrics like accuracy, precision, and recall for distinguishing real vs. fake chip-based cards

< PROJECTS />

FILTER

FEATURED PROJECT

Auto MPG Prediction

This project is a production-grade MLOps pipeline built to predict car fuel efficiency using the UCI Auto MPG dataset. It covers the full ML lifecycle, including data versioning with DVC, experiment tracking with MLflow and DAGsHub, statistical hypothesis testing for model selection, automated testing with pytest, and CI/CD using CircleCI. The final model is served through a FastAPI prediction API, containerized with Docker, and integrated with cloud storage using DigitalOcean Spaces.

Hypothesis testingmachine learningdvcmlopsfastapi