Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
May 20, 2026
·
Manchester NH
Relighting pictures after they are taken
See how AI and ML can relight photos after they are taken. This demo uses open-source models and everyday photography items to show an end-to-end workflow.
Overview
Using AI and ML I built an application that relights pictures after they were taken. I will demo this process producing pictures that are re-lit using off-the-shelf photography items. Examples can be seen here: https://www.instagram.com/jackshap1/.
Video
Links
Personal Instagram profile of Jack Shapiro, featuring 166 posts.
PyTorch-based CUDA photo relighting engine integrating RMBG-2.0 and Depth-Anything-3.
Tech stack
- FluxFlux is the CNCF Graduated project: the definitive GitOps toolkit for continuous, progressive delivery on Kubernetes.Flux is the GitOps engine that automates your application deployment and cluster configuration, ensuring your Kubernetes cluster state always matches the source repository. This CNCF Graduated project integrates directly with your existing infrastructure (GitHub, GitLab, Bitbucket) and OCI artifacts. It provides a robust, auditable, and traceable path for every change: simply commit to Git and Flux handles the reconciliation. Deploy with confidence, manage complex, multi-tenant environments, and eliminate configuration drift using this proven automation platform.
- Runpod GPUsOn-demand GPU cloud providing instant access to H100s, A100s, and RTX 4090s with per-second billing and zero egress fees.RunPod delivers high-performance compute without the hyperscaler tax or provisioning delays. You can spin up a dedicated instance (from a $0.22/hr RTX 3090 to a $3.99/hr H200) in under 60 seconds across 30+ global regions. The platform handles the heavy lifting with pre-configured Docker templates for PyTorch and Jupyter, while per-second billing ensures you only pay for active cycles. Whether you are scaling a cluster to 10,000 GPUs for distributed training or deploying serverless inference with 200ms cold starts, the infrastructure stays out of your way.
- GitHubHost Git repositories and enable massive-scale collaboration (pull requests, issue tracking) for over 100 million developers.GitHub is the world's dominant web-based platform for Git repository hosting and collaborative software development. Built on Linus Torvalds' Git version control system, the platform facilitates 'social coding' by providing essential tools like pull requests, forking, and issue tracking. It currently serves over 100 million developers, managing a massive ecosystem of public and private codebases. Microsoft acquired the company in 2018 for $7.5 billion, solidifying its role as the central hub for open-source and enterprise-level version control.
- Claude CodeAnthropic's agentic coding tool: Unleash Claude's raw power directly in your terminal or IDE to turn complex, hours-long workflows into a single command.Claude Code is Anthropic’s powerful agentic coding assistant, designed for high-velocity development. It operates natively within your terminal, IDE (VS Code, JetBrains), or via a web interface, allowing you to delegate complex tasks like feature building, bug fixing, and codebase navigation. The agent plans, edits files, executes commands, and creates commits, maintaining awareness of your entire project structure. Internally, Anthropic engineers using Claude Code reported a 67% increase in productivity, demonstrating its capacity to deliver significant gains for Pro and Max plan users.
- Runpod GPU'sRunPod provides on-demand cloud GPUs and serverless compute to spin up AI training, inference, and massive workloads in seconds.RunPod delivers high-performance cloud GPU instances (ranging from cost-effective RTX 4090s to enterprise-grade NVIDIA H100s) with per-second billing and zero friction. Developers get instant access to raw compute through secure, customizable pod containers or auto-scaling serverless APIs. By eliminating complex infrastructure overhead, RunPod lets teams deploy custom LLMs, run real-time inference, and execute massive batch jobs without paying the typical hyperscaler premium.
Compose Email
Sending...
Email preview
Loading recent emails...