# AI Engineer - Model Performance

**Company:** [Fathom - AI Notetaker](https://hotfix.jobs/companies/fathom-ai-notetaker)
**Location:** San Francisco, CA
**Role:** ML Engineering
**Skills:** vLLM, Sglang, Tensorrt-Llm, Quantization, Lora, Qlora, Python, Gpu Profiling, CUDA, Multimodal Models, Ray Serve, Modal, Torchtune, Axolotl, Ms-Swift
**Posted:** 2026-04-30

> Optimizes LLM inference stack for speed, cost, and reliability using serving frameworks, quantization, and GPU tuning. Builds fine-tuning pipelines to accelerate AI team model development for production serving of millions of meetings.

## Job Description

## Responsibilities

- Own inference performance: implement speculative decoding, quantization, serving configuration, GPU selection, batching strategies, cold start mitigation, adapter swapping.
- Optimize for spiky traffic (meetings end in 30-minute blocks) focusing on throughput curves.
- Build fine-tuning pipelines for repeatable infrastructure: from JSONL dataset to optimized model ready for serving (LoRA/QLoRA SFT, distillation, DPO).
- Benchmark quantization across GPU families, evaluate serving frameworks (vLLM vs SGLang).
- Optimize GPU spend and debug production inference issues.

## Requirements

**Hard Skills:**
- Deep experience with LLM serving frameworks (**vLLM**, **SGLang**, **TensorRT-LLM**): tuning attention backends, scheduling, CUDA graph warmup, prefix caching.
- Hands-on quantization: weight vs activation, per-channel vs per-tensor scaling, dynamic vs static.
- Production fine-tuning: **LoRA/QLoRA SFT**, training frameworks (**ms-swift**, **Axolotl**, **torchtune**), data formatting, learning rate schedules.
- Strong **Python** for serving infrastructure, benchmarking, training pipelines.
- GPU profiling and performance analysis (compute, memory bandwidth, scheduling).

**Strong Signals:**
- Cost modeling for GPU infrastructure.
- Multimodal models (audio/vision + LLM).
- **Modal**, **Ray Serve**, serverless GPU platforms.
- Audio processing (codecs, chunking, sample rates).
- Building internal tooling.

## Similar roles

- [CoDesign & NextGen Performance Engineer](https://hotfix.jobs/jobs/2cb728b7-a9ea-44b0-b16a-47364021f7be) - Cerebras Systems - Sunnyvale, CA
- [Research Engineer, Privacy](https://hotfix.jobs/jobs/3220c5d0-0e5f-4869-b133-1ed126b4208a) - OpenAI - San Francisco, CA - $380k – $445k/yr
- [Research Engineer](https://hotfix.jobs/jobs/b4d18018-e64d-4828-a8b4-303225328e17) - Console - San Francisco, CA - $200k – $350k/yr
- [Software Engineer, AI Platform](https://hotfix.jobs/jobs/7ddad000-3e1c-47cb-a45b-31d8e72abc4f) - Notion - San Francisco, CA - $180k – $201k/yr
- [Machine Learning Engineer](https://hotfix.jobs/jobs/97713038-58cb-46c5-a9c7-0e1554e1ba4f) - Liftoff - Remote - $215k – $275k/yr

**Apply:** https://hotfix.jobs/jobs/8d7c1585-d9da-4a4d-b407-138a30f6ef79
**Canonical:** https://hotfix.jobs/jobs/8d7c1585-d9da-4a4d-b407-138a30f6ef79