Lead ML Model Training & Post-Training Architect

Inflection AI, Inc.; Palo Alto, Northern (CA, KY); Hybrid; USD 400,000 - 550,000; Full time

Benefits offered by this job

Robust medical, dental and vision with

401k matching

Flexible time off

Holidays & leave

Equity participation

Visa support for Bay Area

Job summary

Inflection AI, Inc. in the San Francisco Bay Area is seeking a Principal Research Engineer to own the model-improvement loop from data and training through evals, post-training, release criteria and production feedback.

This hands-on technical leader will drive training strategies, architecture decisions, and large-scale distributed training across GPUs to ship models that are measurably better for users.

Qualifications

• Experience leading large-scale LLM or foundation-model training or post-training programs.
• Strong experience with transformer-based models and distributed training systems.
• Experience operating large-scale training infrastructure, e.g., GPU clusters.

Responsibilities

• Own the model-improvement roadmap across capability, reliability, and enterprise readiness.
• Lead training and post-training strategy, including supervised fine-tuning and reward modeling.
• Drive architecture and optimization decisions for training and inference.
• Coordinate large-scale training on distributed GPU clusters (thousands of GPUs).
• Define data strategy and evaluation data for production use and feedback loops.
• Build and improve evaluation and release-quality systems and model-readiness reviews.
• Partner with infrastructure and research teams to improve reliability and cost-performance.
• Debug and improve model behavior across data, training, and production.

Skills

Transformer models

Distributed training

Post-training methods

RLHF

SFT

DPO

GRPO

RLAIF

Model evaluation

Budgeting cost and throughput

Education

PhD in Computer Science / ML / AI or related field

Tools

PyTorch

TensorFlow

Job description

Inflection AI, Inc. in the San Francisco Bay Area is seeking a Principal Research Engineer to own the model-improvement loop from data and training through evals, post-training, release criteria and production feedback.

This hands-on technical leader will drive training strategies, architecture decisions, and large-scale distributed training across GPUs to ship models that are measurably better for users.