Staff Deep Learning Engineer – Custom Embedding Architectures

Opportunity Overview

Location: 100% Remote (United States)
Employment Type: Full-Time Direct-Hire via Accel DNA
Compensation Range: $195,000 – $235,000 Base + Equity
Hiring Agency: Accel DNA – Premier Frontier Data & AI Talent Partner

Train, fine-tune, and compress domain-specific dense embedding and reranker models achieving state-of-the-art retrieval performance on specialized corporate data.

Key Responsibilities:

• Train custom bi-encoder and cross-encoder architectures using contrastive learning (InfoNCE, MultipleNegativesRankingLoss).
• Optimize embedding dimensions via Matryoshka Representation Learning (MRL) for efficient multi-scale search indexing.
• Curate hard negative mining pipelines and domain-specific query-document evaluation sets.
• Quantize and compile embedding models into ONNX/TensorRT for sub-10ms CPU and GPU inference.

What You Bring:

• 6+ years of deep learning experience specializing in representation learning and information retrieval.
• Extensive experience with Sentence Transformers, Hugging Face, PyTorch, and contrastive training techniques.
• Familiarity with MTEB (Massive Text Embedding Benchmark) evaluation protocols and custom benchmark creation.
• Strong publication record or demonstrable production retrieval gains.

How to Apply:

Interested in exploring this opportunity? Submit your profile or portfolio directly to contact@accel-dna.com or apply through https://accel-dna.com/contact-us/.