Scaling Distributed AI Workloads on the Spike-A Supercomputer (AI-SIB)
How the AI Supercomputing Initiative Brabant (AI-SIB) and TU/e's Spike-A supercomputer enable scalable multi-GPU model training and code optimization.
Read Technical Summary
Machine Learning Engineer // Supercomputing Brabant (TU/e)
Hi, I'm Siddharth Ganjoo
Machine Learning Engineer & Data Scientist
Scaling distributed AI and high-performance deep learning on the Spike-A supercomputer at Supercomputing Brabant (TU Eindhoven). Specializing in GPU cluster optimization, distributed training pipelines (PyTorch, SLURM, CUDA), and production computer vision.
Background
About & Focus
High-performance computing research and applied AI engineering
Engineering Focus
I am a Machine Learning Engineer and Data Scientist with an MSc in Computer Science & Engineering from Eindhoven University of Technology (TU/e). I specialize in distributed deep learning, high-performance computing on GPU clusters, and production computer vision.
At Supercomputing Brabant (TU Eindhoven), I work with research teams, startups, and high-tech partners to optimize and scale AI workloads on the Spike-A supercomputer. We lower the barrier to advanced supercomputing by migrating cloud bottlenecks to sovereign, distributed European compute.
Previously, I served as Co-Founder & CTO at Loopit.ai (directing computer vision models for vehicle inspection, earning the €1,000 EAISI AI Incentive Award), and engineered telemetry analytics pipelines at Vencomatic Group.
Education
MSc Computer Science & Engineering
Eindhoven University of Technology (TU/e) • 2023 – 2025
Thesis: Reinforcement Learning for Adaptive Sorting (Vencomatic Group)
BTech Computer Engineering
Jamia Millia Islamia University • 2019 – 2023
Grade: 8.54/10 • New Delhi, India Current Initiative TU Eindhoven
Supercomputing Brabant // Meet the Team
View Full Photo With colleagues representing Supercomputing Brabant at our regional ecosystem launch event at TU Eindhoven. We support researchers, deep-tech startups, and industry across North Brabant in scaling AI workloads on the Spike-A supercomputer.
Technical Expertise
High-Performance AI & HPC
Machine Learning & Vision
Data & Cloud Infrastructure
Career & Milestones
Experience
Engineering appointments, research posts, and startup milestones
Eindhoven, Netherlands
Meet the Team: Supercomputing Brabant Event
Supercomputing Brabant // TU Eindhoven
Enlarge Represented Supercomputing Brabant at our regional ecosystem event. Connected with AI founders, researchers, and tech industry leaders to showcase the Spike-A supercomputer cluster, demonstrate scalable GPU workloads, and initiate collaborative supercomputing projects across the Brainport innovation corridor.
Supercomputing BrabantSpike-A ClusterAI OutreachHigh-Tech EcosystemTU Eindhoven
Eindhoven, Netherlands
Machine Learning Engineer — Spike-A
Supercomputing Brabant (AI-SIB) // TU/e
Accelerating and scaling high-performance AI workloads on TU/e's Spike-A supercomputer as part of the regional Supercomputing Brabant initiative (TU/e, Tilburg University, Province of North Brabant, BOM). Supporting research and industry teams in deploying distributed deep learning models, multi-GPU optimization, and transitioning to sovereign European supercomputing infrastructure.
Supercomputing BrabantSpike-A SupercomputerDistributed TrainingPyTorchHPCCUDASLURM
Eersel, Netherlands
Junior Data Scientist
Vencomatic Group
Built automated data pipelines and interactive dashboards using Python, SQL Server, and Streamlit to analyze poultry production telemetry, track KPIs, and detect sensor anomalies on Azure.
PythonSQL ServerStreamlitAzureTelemetry
Eindhoven, Netherlands
MSc Computer Science Graduation
Eindhoven University of Technology (TU/e)
Enlarge Graduated from TU/e with Master's thesis on reinforcement learning for adaptive egg-sorting in collaboration with Vencomatic Group.
Eersel, Netherlands
Master's Thesis Intern
Vencomatic Group
Developed RL-based calibration system reducing false negatives by over 50% (~€25,000 annual savings, ~15,000 eggs saved, ~1,440kg CO₂e avoided). Integrated with knowledge graphs.
Reinforcement LearningKnowledge GraphsOptimization
Evoluon, Eindhoven
AI Summit Brainport 2024
Loopit.ai Showcase
Enlarge Showcased Loopit.ai's platform at the Evoluon to key stakeholders across insurance and automotive industries.
Williamsburg, VA, United States
Research Intern
William & Mary University
Optimized LLM prompts for code generation using CodeT5 by adapting gisting and prompt masking. Benchmarked on CodexGlue with CrystalBLEU and Pass@k.
LLMsCodeT5Prompt GistingNLP
Barcelona, Spain
IESE Ignite Bootcamp Barcelona
EuroTeQ & IESE Business School
Enlarge Represented TU/e in EuroTeQ cohort, accelerating Loopit.ai through mentorship, business development, and VC presentations.
Eindhoven, Netherlands
EAISI AI Incentive Award Winner
Eindhoven AI Systems Institute
Enlarge Won the €1,000 EAISI AI Incentive Award for Loopit.ai's promising automated vehicular damage detection platform.
Eindhoven, Netherlands
Co-Founder & CTO
Loopit.ai
Led end-to-end development of AI-driven platform automating vehicle damage detection and repair estimation. Managed 12-member team. Won €1,000 EAISI AI Incentive Award.
YOLOv8Computer VisionFastAPIDocker
Eindhoven, Netherlands
Philips Innovation Initiative
Philips
Enlarge Participated in Philips innovation initiative, collaborating on cutting-edge technology challenges in the Eindhoven ecosystem.
Eindhoven, Netherlands
Featured in Eindhovens Dagblad
Regional Dutch Press
Enlarge Cashless festival bar system developed for student community featured in the regional newspaper.
Wuppertal, Germany
Research Intern
Bergische Universität Wuppertal
Developed and fine-tuned deep learning models utilizing reinforcement learning and job-shop scheduling to optimize aerospace production in the AlphaMES project.
Deep LearningReinforcement LearningAerospace
India
Software Engineering Intern
Ingersoll Rand
Developed Python-based dashboard with real-time telemetry data integration and modular pipelines for machine usage monitoring and automated reporting on cloud platforms.
PythonDashboardsTelemetry
Portfolio
Selected Projects
Research and commercial machine learning systems driving measurable impact
Startup & Commercial AI
Loopit.ai - Vehicle Damage Detection
AI-driven platform automating vehicular damage detection, localization, and repair estimation for insurance companies. Won the €1,000 EAISI AI Incentive Award.
🏆 Award Winner👥 12-person team⚡ Real-time YOLOv8
AI-SIB // Supercomputing
Spike-A AI Supercomputing Workloads
High-performance AI model acceleration, distributed training pipelines, and multi-GPU cluster scaling on TU/e's Spike-A supercomputer as part of the regional AI-SIB initiative.
💻 Spike-A Cluster⚡ Distributed AI🌐 AI-SIB Brabant
Applied RL Research
RL-Based Egg Sorting Calibration
Reinforcement learning framework reducing false negatives by 50% in high-throughput automated sorting lines at Vencomatic Group.
€25K savings15K eggs saved/yr
LLM Research
CodeT5 Prompt Compression
Optimized prompt gisting and attention masking for code generation models at William & Mary University.
20-40% token savingsPass@k benchmarked
Edge AI
Vehicle Proximity Alert
Real-time YOLOv4-tiny collision warning system detecting nearby vehicles and pedestrians with audio warnings.
30+ FPS edge AIAudio alerts
Document AI
Handwritten Form Digitization
Automated document parsing pipeline extracting handwritten form fields using custom CNNs and Tesseract OCR.
High-accuracy OCRAutomated parsing
Interactive Lab
Spike-A & Neuromorphic AI Simulator
Experience real-time Spiking Neural Network (SNN) dynamics, event-driven synaptic firing, and high-performance compute efficiency benchmarks modeled after our research at TU/e.
Neuromorphic SNN Firing Graph
Energy / Synapse
1.8 pJ -96% vs DNN
Ultra-low neuromorphic power Inference Latency
1.24 ms
Event-driven asynchronous compute Synaptic Event Rate
3,420 ev/s
Distributed GPU execution Total Fired Spikes
0
Live synaptic pulse accumulator Research & Writing
Technical Articles & Case Studies
Deep dives into supercomputing algorithms, Spiking Neural Networks, and industrial machine learning
Calibrating Industrial High-Speed Sorters Using Reinforcement Learning
Reducing false negative sorting errors by 50% through policy optimization and telemetry analysis at Vencomatic Group.
Read Technical Summary
Testimonials
Recommendations
Endorsements from professors and industry collaborators
"Siddharth is amazing. He's smart, humble and he gets things done."
Yanja Dajsuren
Assistant Professor, TU/e
"Siddharth combines technical excellence with practical business acumen."
Gert Guri
CTO, Loopit.ai
"A talented researcher with strong analytical and problem-solving skills."
Bert Sadowski
Associate Professor, TU/e
"Siddharth's dedication to quality and innovation is exceptional."
Alexander Serebrenik
Full Professor, TU/e
Get In Touch
Send a Message
Interested in collaborating, discussing high-performance ML systems, or exploring full-time opportunities? Send a message below.
or connect directly