// Technical Lead · AI Systems Architect · Ph.D. Roboticist

Hi, I'm G Singh

Architecting production-grade AI systems, real-time multimodal observers, and encrypted edge telemetry — taking complex systems from concept to live deployment.

View Projects ↓ Get in Touch
0
Years Experience
0
Lines of Code
0
Repositories
0
Countries Worked

Who I Am

Engineer. Architect. Researcher. Builder.

Technical Lead & AI Systems Architect combining deep research with 10+ years of building high-concurrency backends, real-time vision pipelines, and production AI applications across 5 countries (France, Germany, UK, Portugal, India).

Ph.D. in Robotics from ISAE-Supaéro, France. Creator of real-time multimodal AI observers ("8D-Cogni"), agentic offline AI ("Moon App"), and encrypted edge-to-cloud drone teleoperation pipelines ("8D-Hyphae").

Currently driving technical strategy, engineering culture, and live client deployments as Technical Lead at Eighth Dimension, Germany — leveraging Rust, Python, PyTorch, C, GCP (Vertex AI, Cloud Run, GKE), and TPM 2.0 hardware security.

🏗️
AI & Systems Architect
Low-latency backends in Rust, C, Python & PostgreSQL
🧠
Agentic & Multimodal AI
Multimodal observers (Gemini 2.0), Local RAG, Whisper.cpp
🛸
Edge & Drone Telemetry
Encrypted drone pipelines (AES-256, MAVLink, Jetson ARM64)
☁️
Cloud & Security
GCP ecosystem, GKE, Cloud Run, TPM 2.0 Hardware Binding

Skills & Technologies

The tools and technologies I use to build production systems.

💻
Programming Languages
Rust C / C++ Python JavaScript / TS C# Java SQL
🔧
Frameworks & Tools
React Node.js Electron FastAPI GStreamer Docker Redis PostgreSQL DocuSign API MCP / APIs
🧠
AI & Computer Vision
Gemini 2.0 Claude Opus PyTorch YOLOv8 OpenCV TensorRT SAM 2.1 Local RAG Whisper.cpp ROS / ROS2 CUDA
☁️
Cloud & DevOps
Kubernetes (k8s) GKE (Google Kubernetes Engine) Google Cloud Run GCP Vertex AI Docker Compose CI/CD Linux Edge ARM64 Jetson
🔐
Security & Protocols
TPM 2.0 Hardware Binding AES-256-GCM MAVLink RTSP WebSockets Cryptographic Leases
🎯
Soft Skills
Leadership Team Management Communication Critical Thinking Mentoring

GitHub Activity

Contribution overview and development metrics.

0
Commits
0
Repositories
0
Lines of Code
0
Languages

Projects

Production systems built from architecture to deployment.

VectorX Production Interface
🎛️
VectorX Platform
Engineered for zero-latency live stream switching
Full-stack web streaming and AI pipeline routing platform. Features Oracle Multi-Agent Intelligence with a two-layer predictive AI system (Physics forecasting + Gemini 2.0 Multi-Agent Council) to reason about live stream data with zero frame drops.
⚡ 12ms Redis Stream ⏱️ Sub-50ms Glass-to-Glass 🤖 Gemini 2.0 Council
React Node.js Redis Streams PostgreSQL Docker WebSockets
Scout Recon Dashboard
📊
Scout App (Recon Dashboard)
AI Discovery & Interactive Document Tracking
An intelligent lead discovery and tracking platform powered by Gemini Vertex API. Features a custom PDF viewer that tracks granular user engagement and dwell telemetry on shared documents.
📄 pdf.js Custom Deck 📈 Dwell Telemetry 🤖 Claude Opus Outreach
React Claude Opus 4.8 pdf.js Beacon API Web Analytics
🌙
Moon App (Lunar Analytics)
Local Desktop AI Assistant
An infinitely capable, 100% offline privacy-first AI desktop assistant for Linux using Electron. Integrates local RAG (Ollama, SQLite Vector DB), native dictation (Whisper.cpp), and a custom heuristic engine for system security.
🛡️ 100% Offline RAG 🔍 Security Monitor 🎙️ Voice & Vision I/O
Electron Ollama SQLite Vector Whisper.cpp Local RAG
8D-Sign Enterprise Interface
✍️
8D-Sign
Enterprise DocuSign Deployment & Internal Workflow Management
Customized, deployed, and managed DocuSign integration configured specifically for internal company use. Streamlines secure e-signature pipelines, automated document routing, and compliant multi-signer audit trails.
✒️ DocuSign API Integration 🏢 Internal Enterprise Config 🔒 Managed E-Sign Pipelines
DocuSign API Node.js React Express Enterprise Security
👁️
Vision8 AI Suite
12 computer vision models in one containerized engine
Containerized 12 production microservices: real-time object detection, multi-object tracking, drone detection, crowd analytics, thermal vision, and automated image quality analysis accelerated with TensorRT.
📦 12 Models Repo ⚡ 60 FPS TensorRT 🎥 RTSP Parallel Inference
Python YOLOv8 SAM 3 TensorRT Redis CUDA
🏷️
Annot8 AI
Cut dataset annotation time in half with SAM 2.1
AI-powered image & video dataset annotation platform cutting labeling time by 50%. Integrates Segment Anything Model (SAM 2.1) for zero-shot click-to-segment polygon creation and keyframe propagation.
🚀 50% Speedup 🎯 SAM 2.1 Masking ⚡ 45ms Polygon Gen
Python Flask SAM 2.1 OpenCV React COCO Format
🎯
AI Actor Tracker
Detect and track targeted actors across live streams
Real-time multi-camera RTSP surveillance system for detecting and tracking actors across live streams. Configurable detection keywords with instant WebSockets alert triggers.
🔍 Prompt-Driven Search ⚡ Sub-Second Alert 📹 Multi-RTSP Tracking
Python YOLO Redis OpenCV FastAPI RTSP
🔐
8D-License
Cryptographic Licensing & Hardware Binding
Python-based cryptographic licensing server deployed on Google Cloud Run. Implements enterprise-grade Cryptographic Leases and banking-grade TPM 2.0 Hardware Binding for 100% offline-valid, machine-locked application execution.
🛡️ TPM 2.0 Hardware Locked 🔑 Cryptographic Leases ☁️ Cloud Run Serverless
Python Cryptography TPM 2.0 Google Cloud Run Hardware Binding
🛰️
8D-Hyphae (Queen Pipeline)
Encrypted drone telemetry and video at edge scale
Secure edge-to-cloud drone pipeline built in Rust and C. Delivers end-to-end AES-256 encrypted video, MAVLink flight controls, and sensor telemetry from autonomous drones with sub-50ms glass-to-glass latency.
🔒 AES-256 Transport 🛸 MAVLink Controls ⚡ Sub-50ms Stream
Rust C GStreamer MAVLink AES-256 Linux Edge
🔒
8D-Agent (Flight Computer Runtime)
Ultra-lightweight edge runtime for flight computers
Ultra-lightweight edge runtime deployed on drone flight computers and IoT hardware. Features native C launcher for sub-10ms boot times and self-extracting offline installers for air-gapped field deployments.
⚡ Sub-10ms C Boot 📦 Air-Gapped Installer 🛩️ ARM64 Jetson Orin
C Docker GStreamer Linux Edge ARM64
🧠
AI Observer
Real-Time Multimodal AI Observers
Real-time visual reasoning agents that connect edge cameras and drones to Google's Vertex AI. Utilizes Gemini 2.0 Multimodal Live API over WebSockets for continuous stateful scene analysis and natural-language watch-instructions.
✨ Gemini Multimodal Live 👁️ Stateful Scene Analysis ⚡ WebSockets Audio/Video
Python Gemini Multimodal Live WebSockets Edge AI Vertex AI

Experience

A journey across 5 countries — India, UK, France, Portugal, Germany.

Jan 2025 – Present
Technical Lead
Eighth Dimension — Germany
  • Owned end-to-end delivery of production AI systems from concept to live deployment
  • Architected "8D-Cogni" real-time multimodal AI observer for bidirectional edge telemetry
  • Built privacy-first agentic AI ("Moon App") using Local RAG, vector DBs, and Whisper.cpp
  • Engineered sub-50ms vision pipelines with PyTorch, OpenCV, and MediaPipe
  • Integrated Claude & Gemini LLMs with deep MCP APIs and low-latency Rust/Postgres backends
  • Deployed encrypted drone telemetry (AES-256, MAVLink) & sub-10ms C runtimes for air-gapped Jetson Orin
  • Architected TPM 2.0 hardware-bound cryptographic licensing & enterprise e-sign pipelines on Cloud Run
  • Built 12 containerized vision microservices (TensorRT 60FPS) & SAM 2.1 dataset annotation platforms
  • Led engineering strategy, balancing deep AI research with rapid startup feature delivery
Feb 2023 – Jun 2025
Research Associate
University of the Bundeswehr — Germany
  • Authored 35k+ lines of low-latency, real-time backend code
  • Built ensemble ML models processing real-time user metrics
  • Achieved high-accuracy (F1 = 0.89) predictive data pipelines
  • Mentored four technical team members to accelerate delivery
Feb 2022 – Oct 2022
PostDoc. Researcher
GREYC — France
  • Engineered multi-point routing algorithms optimizing transit efficiency
  • Led successful high-stakes product demonstrations for key clients
  • Developed multi-layer logical planners for complex system navigation
Apr 2020 – Dec 2020
Computer Vision Researcher
SpaceSense.ai — France
  • Wrote scalable Python pipelines for massive data extraction
  • Researched and deployed advanced radar data processing applications
  • Scaled early-stage product features from concept to production
Mar 2017 – Dec 2017
Lead Researcher
M-ITI Madeira & ISR Lisbon — Portugal
  • Developed physiologically attentive user interface for robot teleoperation
  • Built ML models to analyze operators' cognitive states in real-time
  • Published scientific paper within seven months
Apr 2016 – Feb 2017
Research Engineer
Hi-Tech Robotics Systemz — India
  • Developed architecture for teleoperated tracked robot
  • Worked on autonomous shuttle NOVUS DRIVE
  • Obstacle detection with LiDAR and 3D point cloud processing
May 2013 – Mar 2016
Software Engineer
IAP Company / Webmob — India
  • Engineered complex and scalable business applications in C# & .NET
  • Reverse-engineered complex C++ applications for migration

Academic Background

From India to UK to France — a global academic journey.

Doctorate (Ph.D.)
Robotics
ISAE-Supaéro, France
2018 – 2021
Master of Science
Robotics
University of Plymouth, UK
2011 – 2012
Bachelor of Engineering
Information Technology
University of Jammu, India
2005 – 2009

Let's Build Something

Open to collaborations, consulting, and interesting conversations.

Click to open