Established Anno MMXXII  ·  Consulting Engineer

Nikhil Sudhan J

❦
Engineer · Founder · Investigator of Machines
Bangalore, India · B.Tech, C.S.E. 1922–1926* · The Loft Lab

Permit me to introduce myself: a consulting engineer of the modern mechanical sciences, lately occupied with autonomous flying machines, artificial cognition, and the curious art of teaching iron to think. What follows is a faithful dossier of my employments, investigations, and accumulated competencies — set down in plain type, that you may judge their merit for yourself.

I.  Employments

A record of professional engagements

The Loft Lab
Jan 1926 – Present
Founder — Defence-Tech Concern, Koramangala, Bangalore
  • Founded a concern devoted to autonomous unmanned aerial command & control, purpose-built for navigation-denied and high-threat conditions.
Sonata Software
May – Jun 2024
Software Engineer, Apprentice — Chennai
  • Constructed a Flask apparatus joined to Google Cloud Document AI for the automated extraction of documents.
  • Instituted OAuth2 authentication and deployed the works upon a Windows Server machine.
Wipro
May – Jun 2023
Cloud & DevOps Apprentice — Bangalore
  • Containerised and deployed cloud-native applications by means of Docker and Kubernetes upon the Google Cloud Platform.
  • Erected GitOps pipelines for the automated deployment and orderly control of versions.

II.  Investigations

Notable works & contraptions of my own devising

Pluto — An A.I. Integrated UAV Command System
  • A full-stack flying-machine platform: a React frontispiece for the authoring of missions, a Flask engine for the translation of plain instruction into command, and a MAVLink stratum for live telemetry and governance of the craft.
  • Engineered a modular command-engine exposing REST interfaces for real-time execution, with dynamic re-routing and orderly handling of error across navigation-denied flight.
Instruments: Python · Flask · MAVLink · React · REST APIs
Real-Time 3-D Object Tracking upon Edge A.I.
  • Devised a lightweight monocular 3-D perception model by depthwise-separable convolutions; optimised with TensorRT to attain upward of 20 frames per second upon the Jetson Nano, at less than a gigabyte of memory.
  • Benchmarked against full-precision baselines, proving a production-worthy bargain between accuracy and latency, with no dependence whatsoever upon the cloud.
Instruments: PyTorch · TensorRT · KITTI · Jetson Nano · CUDA
An Adaptive Vector Retrieval Engine
  • Architected a cold-and-hot vector storehouse upon FAISS with embeddings summoned on demand; exposed REST interfaces for ingestion, search, and the management of their whole life.
  • Benchmarked retrieval latency and scalability against full in-memory vector repositories, demonstrating superior scale under concurrent floods of enquiry.
Instruments: Python · FAISS · SQLite · REST APIs

III.  Competencies

An inventory of the trade

Languages Python, C, Dart, SQL
Backend & APIs Flask, REST API Design, Microservices, OAuth2, WebSockets
A.I. & Vision YOLO, LoRA, FAISS, Vector Search
Robotics & UAV ROS2, MAVLink, Gazebo, Jetson Nano, Edge Deployment
Cloud & DevOps Docker, Kubernetes, GitOps, CI/CD, GCP, Supabase
Instruments Linux, Git, Fusion 360, Wireshark, Burp Suite, Metasploit

IV.  Certifications

Letters of accreditation

  • Oracle Cloud Infrastructure A.I. Foundations Associate — Oracle, 2025
  • Machine Learning: Natural Language Processing — Udemy
  • Introduction to Networks — Cisco

V.  Correspondence

Whereby I may be reached

— The Calling Card of N. S. Jeeva —
Telephone · +91 75300 92811
LinkedIn Github sudhan.me The Wire (X) Request an Audience
At Your Service