BITS Pilani ACM Student Chapter Logo
BITS Pilani Logo

Machine Learning

Building Intelligent Systems: The Machine Learning Team

The ACM ML Team at BITS Pilani Hyderabad Campus focuses on building data-driven systems that operate reliably in real-world environments. Our work lies at the intersection of statistical learning, computer vision, and intelligent automation. We emphasize strong engineering practices, rigorous evaluation, and an end-to-end understanding of machine learning pipelines - from data collection to deployment.

What We Build

Our projects span vision-based systems, human-computer interaction, and automated decision pipelines. Each project is engineered with reproducibility, scalability, and robustness in mind.

1. Face Anonymizer System

A privacy-preserving framework for intelligent identity obfuscation using generative AI.

  • Leverages Flow Matching Models to generate realistic, anonymized identities, an application of this architecture that has never been done before in real-time.
  • Uses precise Facial Landmark Detection to map geometry and ensure accurate alignment during the anonymization process.
  • Applies advanced Image Warping techniques to seamlessly blend generated features with the source footage, removing visual artifacts.
  • Core Concepts: Computer Vision, Flow Matching, Facial Landmark Detection, Image Warping, Privacy-Preserving AI.

2. Air-Draw: 2D Drawing in Mid-Air

A gesture-driven interface that captures finger motion and renders trajectories digitally, allowing users to paint on a virtual canvas without physical touch.

  • Leverages MediaPipe for real-time 21-point skeletal hand tracking to map physical gestures to digital coordinates.
  • Implements distinct gesture-recognition logic to switch seamlessly between drawing modes and tool selection.
  • Applies Quadratic Bézier smoothing algorithms to eliminate hand jitter and ensure fluid, natural lines.
  • Core Concepts: Computer Vision, Human-Computer Interaction (HCI), Real-Time Tracking, Signal Smoothing.

3. Resume Screening Intelligence System

An AI-powered placement assistant fine-tuned on BPHC campus data to provide personalized career insights and predictions.

  • Leverages a Large Language Model (LLM) fine-tuned on historical campus placement records for context-aware analysis.
  • Evaluates Resume vs. Job Description (JD) alignment to predict interview qualification probability.
  • Identifies the top 5 company matches for a candidate based on successful alumni profiles.
  • Provides actionable feedback on resume improvements tailored to specific target companies.
  • Core Concepts: Natural Language Processing (NLP), Large Language Models (LLMs), Fine-Tuning, Predictive Analytics.

How We Build

We leverage foundational frameworks and libraries to gain a deep, first-principles understanding of AI model development, focusing on reliability and performance.

  • Frameworks: PyTorch, TensorFlow, Scikit-learn
  • Core Architectures: Foundational Neural Networks, Sequential Models, Advanced Computer Vision Architectures
  • Domains: NLP, Computer Vision (CV), Agentic AI, RAG, Responsible AI, Real-Time Inference
  • Libraries: LangChain, NumPy, Pandas, OpenCV, Matplotlib, FastAPI

Why Join the ML Team?

The ACM ML Team is for students who want to develop strong technical depth in machine learning systems engineering. By joining, you will:

  • Design and implement ML pipelines end-to-end
  • Gain experience in dataset construction and model validation
  • Learn to diagnose model failures and optimize systems for production-like constraints
  • Work on interdisciplinary projects spanning vision, language, and automation
  • Build a portfolio that demonstrates engineering rigor rather than surface-level usage

If you’re interested in building reliable, data-driven systems and understanding how learning algorithms behave in practice, this team will give you the platform to do exactly that.