Technical Architect - AI Systems & Platform Internals

Responsibilities

• Design and evolve large-scale AI systems supporting agentic workflows, multimodal models, and research workloads.
• Architect high-throughput, low-latency inference systems across large-scale GPU clusters, optimizing for latency, throughput, and cost.
• Guide engineering decisions regarding GPU kernels, memory movement, and distributed execution using CUDA, Triton, and NCCL/RCCL.
• Design context engineering frameworks, including RAG, long-context management, and dynamic context assembly.
• Build cost optimization frameworks to reduce token usage and infrastructure spend through model routing and semantic caching.
• Collaborate with research teams to support distributed training, orchestration, and fault tolerance.
• Architect validation and release systems to ensure model updates and platform changes are safe, performant, and regression-free.
• Define telemetry, observability, and reliability standards for AI infrastructure.

Requirements

• 10–12 years of experience in software engineering, systems architecture, ML infrastructure, or distributed systems.
• Strong hands-on engineering proficiency in Python and at least one systems language (C++, Go, Rust, Java, or TypeScript).
• Deep understanding of distributed systems, production infrastructure, and fault-tolerant architecture.
• Practical experience with GPU systems, accelerator-based workloads, and CUDA/Triton-style programming.
• Experience with ML frameworks and serving stacks such as PyTorch, JAX, vLLM, or Triton.
• Proven ability to debug complex problems across model behavior, runtime systems, and distributed networking.

Preferred Qualifications

• Experience with LLM inference optimization, including tensor parallelism, pipeline parallelism, and KV-cache management.
• Experience profiling GPU workloads using tools like Nsight Systems, Nsight Compute, or Prometheus.
• Background in designing context engineering platforms or model-routing frameworks.
• Experience with large-scale distributed training and RL infrastructure.

About the Company

Accellor is an AI-native services firm purpose-built for the post-ChatGPT era. We focus on delivering measurable business outcomes through advanced AI, data, and engineering capabilities, helping Fortune 100 enterprises operationalize AI at scale.