About us
Flam is building the next generation of AI-native content formats that transform passive customer touchpoints into humanlike visual interactions. We power enterprises to create immersive, interactive experiences across marketing, communication, and commerce using cutting-edge generative AI.
Core Mission: We believe the most powerful interactions aren't transactional—they're creative, tactile, and alive. Flam is enabling businesses to go "beyond videos" with intelligent, interactive visual content.
Our Technology / AI Models
Fable 2.0
- Purpose: AI-native content generation and visual synthesis
- Technology: Flow matching diffusion transformer model
- Capabilities: Generates native RGBA content with dynamic motion, fine structural detail, and strong foreground presence
- Use Case: Creating immersive visual experiences from text prompts
Fantom 1.0
- Purpose: Identity-preserving video synthesis
- Technology: Advanced video synthesis model
- Capabilities: Creates temporally coherent video streams from a single face image and audio input
- Use Case: Personalized video communications and interactive avatars
Falcon 1.0
- Purpose: Conversational intelligence and personality-aware responses
- Technology: 30B parameter Mixture-of-Experts (MoE) large language model
- Capabilities: Delivers personality-aware and context-adaptive responses
- Use Case: Intelligent conversational layers within Flam experiences
Core Solution Pillars
- Engage (Marketing)- Make Flams part of your marketing mix by integrating interactive visual content across every customer touchpoint, driving higher engagement and brand recall.
- Connect (Communication) - Utilize Flam for personalized, immersive communication experiences that create deeper human-machine interactions and improve customer relationships.
- Convert (Commerce) - Transform visual interactions into measurable business outcomes with integrated checkout and commerce capabilities built into Flam experiences.
Key Technology Features
Interactive Streaming at Scale
- Immersive visuals and interactive controls with zero load time
- First frame loads in under 300 milliseconds
- No app download required (app-less infrastructure)
- Pro-grade immersion quality with humanlike interactions
- Omni-channel presence and integrated checkout
Prompt-Powered Creation
- Cut development weeks down to seconds with AI-driven generation
- Text-to-Flam conversion using natural language prompts
- One-click variant generation for A/B testing
- Integrated editor for fine-tuning (text, image, layers, audio, 3D, motion)
Distribution Channels - Flams can be distributed through multiple channels for maximum reach:
- Web Embed: Embed directly into websites with a simple code snippet
- Sharable Links: Distribute via digital ads, social media, SMS, and email
- Flam Codes: Integrate with physical channels (outdoor ads, CTV, retail stores, events)
What You'll Be Doing:
- Design and implement custom Kubernetes controllers that manage GPU node lifecycle across multiple cloud providers, handling provisioning, bootstrapping, and de provisioning with zero manual intervention
- Build and maintain autoscaling systems that analyze pod resource requirements (GPU count, memory, CPU) and automatically provision appropriate instance types from provider APIs
- Develop microVM-based isolation for multi-tenant GPU workloads using cloud-hypervisor with VFIO GPU passthrough, enabling secure, high-performance compute for inference and training workloads
- Create node bootstrapping automation that provisions bare-metal or cloud instances with container runtimes, GPU drivers, Kubernetes components, and custom networking configurations via cloud-init, Ansible, or Terraform
- Implement sophisticated networking solutions connecting hybrid infrastructure - GKE/GCE control planes with external GPU workers using WireGuard, custom CNI configurations, and cross-cloud service mesh
- Build Kubernetes operators and CRDs for managing ML infrastructure components like model registries, inference endpoints, training job orchestration, and GPU time-slicing configurations
- Design monitoring, cost optimization, and capacity planning systems that provide visibility into GPU utilization, workload patterns, and infrastructure efficiency across heterogeneous compute pools.
- Work closely with ML researchers to understand workload requirements and translate them into infrastructure automation that enables rapid experimentation and production deployment.
What We Need to See :
- 5+ years of experience building production Kubernetes systems with deep expertise in controllers, operators, CustomResourceDefinitions, and API machinery
- Strong proficiency in Go and experience building scalable, reliable services that manage complex distributed systems
- Hands-on experience with GPU infrastructure in Kubernetes - NVIDIA GPU Operator, device plugins, time-slicing configurations, or custom GPU scheduling logic
- Deep understanding of Kubernetes architecture including admission controllers, scheduler extenders, resource lifecycle management, and cluster autoscaling mechanisms
- Demonstrated ability to design and implement automation systems that replace manual processes with API-driven, self-service tooling
- Experience with at least one cloud provider's APIs (GCP, AWS, Azure) for programmatic compute provisioning and management
- Strong Linux systems knowledge including networking (iptables, WireGuard, CNIs), storage (LVM, device mapper), and virtualization (KVM, QEMU, cloud-hypervisor)
- Bachelor's/Master's degree in Computer Science, Engineering, or equivalent practical experience
Ways to Stand Out from the Crowd :
- Experience building hybrid/multi-cloud Kubernetes architectures with cross-provider networking and unified control planes
- Deep familiarity with microVM technologies (Firecracker, cloud-hypervisor, Kata Containers) and their application to GPU workloads
- Hands-on experience with VFIO GPU passthrough, SR-IOV, MIG (Multi-Instance GPU), or other GPU virtualization technologies
- Track record of building custom cloud controllers or provider implementations for bare-metal or specialized compute
- Experience with ML infrastructure patterns - model serving, training orchestration, experiment tracking, or distributed training frameworks
- Contributions to upstream Kubernetes projects, CNCF ecosystem tools, or GPU-related open source projects
- Understanding of cost optimization strategies for GPU compute, including spot instances, preemption handling, and intelligent workload placement
- Experience with Infrastructure as Code (Terraform, Pulumi) for complex multi-provider deployment
Why Join Flam?
- Flam is building the future of AI-native content. We're at the intersection of:
- Generative AI: Cutting-edge models for content creation
- Enterprise Scale: Trusted by Fortune 500 companies
- Product Innovation: Transforming how brands communicate
- Rapid Growth: Scaling new technology to global enterprises