Satyajit Ghana

Experience

• Inkers Technology

  Bengaluru, India
  • Head of EngineeringJan 2024 – Present
  • Deep Learning Software EngineerJun 2022 – Jan 2024
  • Deep Learning AssociateJul 2021 – Jun 2022
  • — Lead engineering for industrial-AI products: 3D perception, LiDAR/point-cloud pipelines, and structural-defect analysis.
  • — Design custom neural networks, CUDA kernels, and high-performance C++/gRPC services for production deployment.
  • — Named inventor on 2 USPTO patents pending (assigned to Inkers); grew from Deep Learning Associate to Head of Engineering.

• The School of A.I.

  Remote
  • MLOps InstructorJan 2022 – Jan 2023
  • — Designed and taught EMLO 2.0 — an end-to-end MLOps course (training, packaging, deployment, monitoring).
  • — Contributed to EVA 4.0, the deep computer-vision program.

Education

• M.S. Ramaiah University of Applied Sciences

  2021

  B.Tech · Bangalore, India · CGPA 9.78/10 · Silver Medalist

Skills

Deep Learning

PyTorch · TensorFlow · Computer Vision · 3D / Point Clouds / LiDAR · GenAI (SDXL, LLMs)

Systems

C++ · C · Rust · CUDA · gRPC · MongoDB

MLOps

Kubernetes · AWS · GCP · Docker

Web

TypeScript · React · Next.js

Selected Projects

• torch-point-ops

  High-performance PyTorch operators for point-cloud and 3D geometry processing.

• PV-LIO-for-HBA

  Point-to-voxel LiDAR-inertial odometry adapted for hierarchical bundle adjustment.

• ige_lio

  Iterated error-state Kalman-filter LiDAR-inertial odometry experiments.

• sdxl-dreambooth-finetune

  DreamBooth fine-tuning pipeline for Stable Diffusion XL subject personalization.

• TSAI-DeepVision-EVA4.0

  Coursework and experiments from The School of AI's EVA 4.0 computer-vision program.

• PadhAI-Course

  Implementations and notes from the PadhAI deep-learning course.

Publications

• Adaptive Visual Learning Using Augmented Reality and Machine Learning Techniques

  Journal of Computational and Theoretical Nanoscience · Vol. 17, No. 11, pp. 4952–4956 · 2020 · DOI 10.1166/jctn.2020.8982

Patents

• Method and System for Performing Structural Defect Analysis in a Structural Environment

  Pending · US 19/634,310 · filed 2026-03-31 · Inkers Technology

• Data Acquisition Device

  Pending · US 19/634,339 · filed 2026-03-31 · Inkers Technology