AI Engineer

About the Engagement

FlareTech is building "Context Intelligence," an AI platform centered on a proprietary "process neural net" architecture. Rather than relying on standard NLP or data warehouses, the platform builds model graphs that reverse-engineer data models directly from the business, enabling clients to inject knowledge management indexed to their own goals and strategy. FlareTech has secured a Proof of Concept with a major global enterprise client (hundreds of thousands of employees) and needs to accelerate delivery toward an October 1 POC milestone, with full production deployment targeted for November/December.

This role owns the "Context" domain specifically — context artifacts, ingestion workers, and fine-tuning — while the existing FlareTech team continues building out process automation and the rest of the platform in parallel.

Main Responsibilities

• Own the Context domain end-to-end: context artifacts, ingestion workers, and the fine-tuning pipeline within the Context Intelligence platform.
• Design, implement, and operate production-grade fine-tuning workflows for localized (on-prem / self-hosted) LLMs.
• Build and maintain ingestion workers that structure and feed data into the model graph architecture.
• Support the reverse-engineering of client data models into model graphs, in line with FlareTech’s process-neural-net approach.
• Deploy and manage local LLM hosting to meet enterprise confidentiality requirements, as an alternative to cloud-hosted LLM APIs.
• Engage directly with Darshan (VP of IT) to scope work, set priorities, and integrate deliverables into the critical path for the October 1 POC.
• Coordinate with the existing platform team to ensure the Context domain integrates cleanly with process automation and other platform areas.
• Prepare technical documentation on the process neural net / model graphing architecture to support client technical due diligence.
• Help accelerate the POC timeline, then transition the Context domain into full production for the November/December deployment.

Required Skills & Experience

• Production experience fine-tuning LLMs — has shipped fine-tuned models into a live product, not only prompt engineering or inference-time RAG.
• Hands-on experience deploying and operating localized / on-prem / self-hosted LLMs (e.g., via vLLM, TGI, Ollama, or comparable serving stacks).
• Strong understanding of Retrieval-Augmented Generation (RAG) architectures, and how fine-tuning and RAG complement each other.
• Experience building data ingestion pipelines — workers that clean, structure, and prepare data for model consumption.
• Familiarity with graph-based or structured knowledge representations (model graphs, knowledge graphs), including reverse-engineering data models from unstructured or semi-structured sources.
• Comfortable operating in a fast-moving, ambiguous R&D environment against a hard external deadline (client POC).
• Strong communication skills — will scope work directly with a client-side VP of IT with minimal intermediary process.
• A security- and confidentiality-first mindset, understanding why enterprise clients require local LLM deployment over cloud APIs.

Nice to Have

• Experience building no-code or low-code AI/workflow automation platforms.
• Exposure to agentic frameworks or multi-agent systems (useful context, even though this platform favors fine-tuning over complex agent authoring).
• Background in enterprise knowledge management or business process automation.
• Experience with platform/marketplace business models involving third-party extensibility.