Skip to content

Sidhardh SAI/ML Engineer & Researcher

I design intelligent systems that learn from structure.

Final-year Integrated M.Tech student building research-grade models across graph neural networks, state space models, and generative AI.

[01]About

Built on rigor, driven by curiosity.

I am a final-year Integrated M.Tech (Software Engineering) student at VIT, focused on machine learning and deep learning research. I am a co-author on IEEE-targeted papers, and I care about the full arc of a project — from dataset construction through model design, evaluation, and ablation.

My work spans graph neural networks, state space models, and generative AI. I like problems where the structure of the data is the hard part, and I enjoy turning a vague research idea into a reproducible experiment with results I can defend.

0+
Years exploring ML
0
IEEE-targeted papers
0+
Models trained & ablated
0
Core research areas

Focus areas

Graph Neural Networks

01

Message passing, graph transformers, and representation learning on relational and molecular data.

State Space Models

02

Long-sequence modeling with structured state spaces (S4 / Mamba-style) as efficient alternatives to attention.

Generative AI

03

Diffusion and latent generative models for synthesis, augmentation, and controllable generation.

Evaluation & Ablation

04

Reproducible pipelines, honest baselines, and ablations that isolate what actually drives performance.

[02]Selected Work

Selected work.

A few projects that show how I think — from data to model to results. Code is on GitHub where it can be shared.

A hybrid that injects state-space sequence mixing into message passing, targeting long-range dependencies that standard GNNs miss.

  • PyTorch
  • PyG
  • State Space Models
  • GNN
2025 Code

Graph neural network for molecular property prediction with a reproducible benchmark suite and full ablation harness.

  • PyTorch Geometric
  • Cheminformatics
  • GNN
2024 Code

Conditional latent diffusion pipeline for dataset augmentation, with classifier-free guidance and quality/diversity metrics.

  • Diffusion
  • Generative AI
  • PyTorch
2024 Code

Lightweight experiment framework for sweeping, logging, and ablating deep-learning runs without the boilerplate.

  • Python
  • MLOps
  • Weights & Biases
2023 Code
[03]Research & Publications

Research & publications.

Peer-review-targeted work. Co-authored unless noted otherwise.

  • 01

    Structured State Spaces for Long-Range Reasoning on Graphs

    S. S. et al.

    IEEE (target venue)2025Under review
  • 02

    Ablation-First Evaluation of Message-Passing Architectures

    S. S. et al.

    IEEE (target venue)2025In preparation
[04]Skills & Stack

Skills & stack.

The tools I reach for, grouped by what they do.

01ML Frameworks

  • PyTorch
  • PyTorch Geometric
  • TensorFlow
  • JAX
  • Hugging Face
  • scikit-learn

02Infra & MLOps

  • CUDA
  • Docker
  • Weights & Biases
  • MLflow
  • DVC
  • Linux
  • Git

03Languages

  • Python
  • C++
  • TypeScript
  • SQL

04Data & Tooling

  • NumPy
  • Pandas
  • Matplotlib
  • Jupyter
  • NetworkX
  • RDKit
[05]Contact

Let's build something intelligent.

Open to research collaborations, ML engineering roles, and interesting problems. The fastest way to reach me is email.

Available for 2026 roles & research collaborations