Gen AI architect

Prodapt; San Francisco (CA); On-site; USD 150,000 - 190,000; Full time

Job summary

A leading AI consulting firm in San Francisco is seeking an Applied AI Engineer to optimize models and provide technical vision. Responsibilities include fine-tuning models using advanced techniques, performing quantization for cost efficiency, and acting as a strategic consultant to executives. The ideal candidate has strong leadership skills and expertise in model optimization, capable of influencing C-level clients and defining the AI culture.

Qualifications

• Proven experience in fine-tuning models like Llama 3 and Mistral.
• Knowledge of GPU memory management and quantization techniques.
• Ability to influence C-level executives.

Responsibilities

• Implement PEFT to adapt models to specific domains.
• Perform model quantization to reduce inference costs.
• Continuously research the latest AI advancements.
• Provide strategic consulting and define the AI architecture vision.

Skills

Applied AI engineering

Model Fine-Tuning

Optimization & Quantization

Strategic Consulting

Executive communication

Cross-org leadership

Tools

vLLM

TGI

GGUF / AWQ

Llama 3 / Mistral

Job description

Role Summary

You represent the pinnacle of Applied AI engineering. You are not just using APIs; you are optimizing the models themselves. You understand the mathematics behind the attention mechanism, you know how to squeeze performance out of GPUs, and you can customize models for specific domains. You provide the high‑level technical vision and handle the most difficult edge cases.

Key Responsibilities

• Model Fine‑Tuning: Implement PEFT (Parameter‑Efficient Fine‑Tuning), LoRA, and QLoRA to adapt open‑source models (Llama 3, Mistral) to specific client domains.
• Optimization & Quantization: Perform model quantization to reduce inference costs and latency without sacrificing quality. Manage Dense Vectors and embedding optimizations.
• State‑of‑the‑Art Exploration: Continuously research and implement the latest advancements (e.g., State Space Models, Long‑Context optimizations) into client deliverables.
• Strategic Consulting: Act as a trusted advisor to C‑level client executives, defining the "Art of the Possible" and guiding long‑term AI roadmaps.

Technical Requirements

• Model Ops: Serving custom models (vLLM, TGI), GPU memory management, Quantization techniques (GGUF, AWQ).
• Advanced Data: Training data curation, synthetic data generation, RLHF concepts.
• Tech Leadership: Ability to define the technical culture and set standards for the entire FDE organization.

Soft Skills

• Executive communication and ability to influence C‑level leaders.
• Thought leadership and industry presence (conferences, playbooks, forums).
• Cross‑org leadership and conflict resolution.
• Ability to define long‑term AI vision and cultural standards.
• Strategic decision‑making balancing cost, risk, and performance.