Fernando Wallick cover
Fernando Wallick
Fernando Wallick

Senior Distributed AI Systems Architect

Actively looking · Member since 15 Nov 2026
Location
Jávea, Spain
Desired salary
60–100 EUR/hourly
Work preference
Remote Only / Full Time, Part Time, Contract, Freelance, Temporary
Experience level
Senior

About

Professional summary

I am a hands-on software architect with more than 24 years of experience building low-latency, memory-safe distributed systems in Rust, C++, Go, and Python. I specialize in systems that operate at large scale across cloud, edge, gaming, automotive, and defense environments.

I design distributed AI platforms spanning model training, PEFT, inference, agent systems, and GPU infrastructure. My work includes NVIDIA AI factories using NeMo, Megatron, TensorRT-LLM, NIM, KServe, Triton, and Kubernetes-based orchestration.

I build zero-copy data engineering platforms using the modern Apache and Rust ecosystem, including Iggy, Redpanda, Polars, DataFusion, Ballista, Arrow Flight, Parquet, and Iceberg. I focus on high-throughput analytics, streaming systems, lakehouse architectures, and efficient data movement.

I have extensive cloud-native and platform engineering expertise across AWS, GCP, Cloudflare, Kubernetes, Crossplane, OpenTofu, GitOps, and service mesh technologies. I create reproducible infrastructure and self-healing delivery workflows for complex distributed environments.

My systems have supported hundreds of millions of user accounts and significant recurring revenue while maintaining very low downtime. I am available for fully remote contract engagements worldwide, especially for distributed AI, cloud/edge infrastructure, data engineering, and high-performance systems challenges.

Notice period: 2 weeks

Skills

43 capabilities

Tech stack & tools

Working toolkit

Experience

Career history

Senior Cloud/Edge Distributed AI Systems Architect DroneGrid with VW/Audi Automotive

I architect cloud and edge distributed AI systems for physical AI self-driving UGV and UAV fleets. I built the Legio P2P framework using C++20, Rust, Cap'n Proto RPC, QUIC, TCP, ROS2/DDS, libp2p, Media-over-QUIC, and CRDT fleet state, achieving p99 latency below 8 ms and p999 below 20 ms.

I designed Kubernetes-based GPU inference and AI infrastructure across GCP GKE, MicroK8s, KubeEdge, Knative, and KEDA. A Rust kube-rs operator automates deployment of the NVIDIA GPU Operator, KServe, Triton, KAgent, AgentGateway, Rust Rig agents, and Qdrant RAG.

I also developed a Rust analytics and model-improvement platform with DataFusion Ballista, Arrow Flight, Polars, Leptos, Axum, and async-graphql. The platform feeds prediction-loss signals into QLoRA PEFT retuning and supports NVIDIA Omniverse/OpenUSD digital-twin simulation.

Senior Distributed AI Systems Architect / Senior AWS Cloud and Cloudflare Edge AI Factory Architect Activision

I architected high-scale game asset auction, exchange, AI, and cloud platforms supporting 250M+ player accounts and $225M MRR. I built zero-copy, memory-safe C++20 and Rust transaction services with Go gRPC, GraphQL, and Cap'n Proto edge and control-plane microservices across AWS EKS, EKS Fargate, and Cloudflare Workers.

I built an end-to-end NVIDIA AI factory for auction-fairness and pricing models using NeMo, Megatron-Core, NCCL, Kubeflow, Kueue, TensorRT-LLM, NIM, KServe, vLLM, SGLang, Triton, and KEDA. The platform reduced inference cost per request by 3.4x compared with the prior Python serving path.

I designed a Rust-native analytics lakehouse using Apache Iggy, Polars, DataFusion, Ballista, Arrow Flight, Parquet, Iceberg, Cloudflare R2, Aurora PostgreSQL, and ElastiCache Redis. I also codified delivery through kube-rs, Crossplane, OpenTofu, Flux, Argo CD, Cilium, Istio, and Vault.

Senior Distributed AI Systems Architect / Senior AWS Cloud and Cloudflare Edge AI Platform Architect Activision

I grew the P2P player-to-player inventory asset exchange platform to $200M MRR across 250M+ accounts. I developed Rust Tokio WebSocket services, Apache Iggy topic clusters, Cap'n Proto RPC over QUIC, Go gRPC and GraphQL gateways, and regional HA services using Aurora PostgreSQL and Redshift Serverless.

I established GPU training pipelines for auction and exchange reinforcement-learning policies using PyTorch, Megatron-LM, NCCL, NeMo, Kubeflow, Volcano, Rust Burn, and Candle. Models were served through Triton and KServe on EKS and Knative.

I introduced Crossplane, OpenTofu, Flux tofu-controller, Argo CD, Tekton, Argo Events, and Argo Workflows as the platform delivery backbone. I also designed the Arrow-native data plane based on Iggy, DataFusion, Ballista, Polars, Parquet, Iceberg, and Cloudflare R2.

Senior C++/Rust Cloud/Edge AI Distributed Systems Architect NVIDIA

I worked on Raytheon DoD machine-vision UAV programs, building distributed cloud and edge AI systems for tactical fleets operating over lossy and intermittently offline links.

I developed a P2P CRDT GPU streaming-SQL swarm engine in C++ and Rust, incorporating CUDA kernels, Raft replication, UDP P2P sockets, DDS, and DNN-RL query planning. The platform ran on GCP GKE GPU nodes and Anthos.

Senior C++ Distributed Systems Architect Activision

I developed hybrid WAN-LAN P2P matchmaking systems for PlayStation 3 and PC AAA titles. The platform supported more than 200M accounts and approximately $175M MRR.

I built distributed CSPRNG ranking, party, and playlist services with ChaCha20, deterministic-lockstep P2P networking in C++11, Boost.Asio reliable UDP, DDS, Redis sorted sets, spatial partitioning, AWS EC2 Auto Scaling, Route 53 latency routing, and HAProxy.

Senior C++ P2P AI Distributed GPU Systems Architect / Senior C/C++ Linux Kernel Storage Software Eng NVIDIA

I worked on Raytheon DoD machine-vision UAV programs, developing distributed GPU systems and storage software. I built the original CUDA C++ PostgreSQL-extension GPU streaming-SQL swarm engine and a P2P CRDT data synchronization platform.

My work included OpenCV, DDS, UDP networking, Tesla GPU systems, DNN machine-vision drone workloads, and Linux kernel storage engineering.

Education

Learning history

Stuttgart University

M.S., Computer Science and Technology

Stuttgart University

B.S., Computer Science and Technology

Legio P2P-PGP-RS Rust Networking Substrate

Legio P2P-PGP-RS Rust Networking Substrate

Legio is the Rust-native networking substrate under every DroneGrid product: a 30-crate workspace that gives UGV, UAV, SBC and digital-twin fleets one CRDT-converged state over…

View project
Vesuvius-RS MCP Server

Vesuvius-RS MCP Server

Vesuvius is the highest-quality Rust-native Model Context Protocol server. Every span that flows to or from your tools is judged by compiled, blake3-sealed sentinel policies…

View project
Vigilantus-rs MCP Server Context Sentinel Service

Vigilantus-rs MCP Server Context Sentinel Service

Vigilantus is a Rust-native adapter that sits in front of the MCP server you already run — any language, any stack, untouched — and removes…

View project

This professional hasn’t listed any services yet.

Jobs Talent AI Tools Salaries
Menu