Senior/Principal Local LLM & Generative AI Platform Engineer
Remote • US • Full Time
Posted 1mo ago
Job Location
US
Tech Stack
Remote Work Policy
Fully remote
Employment Type
Full Time
Categories
LLM Engineer
About the job
Parallel Wireless is seeking a hands-on technical leader to build and operate a secure local large-language-model (LLM) platform. This platform will enable engineering and business teams to leverage generative AI with proprietary company data, including source code, product documentation, and technical standards, all within a secure, company-controlled environment. This senior individual contributor role involves applied LLM engineering, platform architecture, data pipelines, security, and production operations. You will transform prototypes into a dependable internal capability by selecting and optimizing open-weight models, building permission-aware retrieval systems, creating reusable APIs and tools, integrating with existing workflows, and establishing metrics for quality, safety, latency, capacity, and business value. The role emphasizes building a trustworthy enterprise LLM solution that includes robust source grounding, access controls, repeatable evaluation, observable services, and a sustainable operating model.
Responsibilities
- Own the architecture and technical roadmap for a secure, reliable, and maintainable local LLM platform.
- Partner with various teams to prioritize use cases and translate them into platform requirements.
- Build a modular inference and model-gateway layer with stable APIs and flexible model management.
- Evaluate and document open-weight language, code, embedding, and multimodal models.
- Optimize model serving across various hardware resources using techniques like quantization and caching.
- Design and operate RAG and enterprise-search pipelines for various data sources.
- Enforce source-system permissions throughout ingestion and retrieval.
- Establish evaluation datasets and automated evaluation for retrieval quality and accuracy.
- Create release gates and reproducible regression tests for platform changes.
- Implement end-to-end observability for model and agent workflows.
- Design safe tool-calling and agent workflows with least-privilege access and sandboxing.
- Integrate the platform into existing employee tools through SDKs and APIs.
- Build operational foundations for production use, including CI/CD, backups, and incident response.
- Protect proprietary and personal information through security measures and defenses against various risks.
- Determine when fine-tuning or other model adaptation is justified by measured quality gains.
- Make the platform usable beyond the core AI team through documentation and training.
Requirements
- Hands-on technical leadership experience.
- Experience building and operating secure local LLM platforms.
- Proficiency in applied LLM engineering, platform architecture, and production operations.
- Experience selecting and optimizing open-weight models.
- Experience building permission-aware retrieval systems.
- Experience creating reusable APIs and tools.
- Experience integrating with existing engineering workflows.
- Understanding of enterprise LLM requirements including source grounding, access controls, evaluation, and observability.
- Experience with search, retrieval-augmented generation (RAG), prompting, workflow automation, or model adaptation.
- Experience with model serving optimization techniques (e.g., continuous batching, caching, parallelism, quantization).
- Experience designing and operating RAG and enterprise-search pipelines.
- Experience enforcing source-system permissions and integrating with identity and access control systems.
- Experience establishing evaluation datasets and automated evaluation for LLM performance.
- Experience with release gates, reproducible regression tests, and CI/CD.
- Experience implementing end-to-end observability for AI workflows.
- Experience designing safe tool-calling and agent workflows.
- Experience protecting sensitive information through security measures.
- Experience with parameter-efficient fine-tuning or other model adaptation techniques.