Skip to main content
TALENT247 // SPECIALIST AI EXECUTIVE SEARCH & TECHNICAL RECRUITMENT
ML · Machine Learning

Staff Neural Retrieval & Search Ranking ML Engineer

San Francisco, CA · Full-time · On-site

AI Specialism

Machine Learning

Workplace

On-site

Location

San Francisco, CA

Compensation

$255k – $355k

annual package

Role Dossier & Specification

REF: STAFF-NEURAL-RETRI

THE CLIENT & MANDATE

Our client is a multi-billion-dollar real-time AI answer and deep-research engine headquartered in San Francisco, serving billions of grounded, citation-backed queries every month across consumer, enterprise, and developer APIs. Unlike static pretrained LLMs whose world knowledge freezes at training cutoff, their core technical moat is a planetary-scale, sub-100ms neural web search and multi-stage ranking stack built specifically for LLM reasoning agents rather than ten-blue-link human clicks.

They have retained Talent247 to appoint a Staff Neural Retrieval & Search Ranking ML Engineer onsite in San Francisco, CA. Serving as a technical lead within their Core Search & Retrieval ML organization, the successful candidate will architect the next generation of dense, learned-sparse, late-interaction, and cross-encoder ranking models over hundreds of billions of live web and enterprise documents.

KEY RESPONSIBILITIES

  • Define the architectural roadmap and training recipes for multi-stage neural search retrieval: multilingual dense embeddings, learned sparse/lexical expansion, token-level late interaction, and ultra-low-latency cross-encoder rerankers.
  • Optimize retrieval objective functions specifically for downstream LLM reasoning utility—maximizing factual completeness, primary-source authority, temporal freshness, and multi-hop evidence diversity rather than clickbait CTR.
  • Co-design quantized embedding formats (binary/INT4/Matryoshka representations) and GPU/CPU approximate nearest-neighbor (HNSW/IVF-PQ/DiskANN) pruning cascades alongside C++/Rust search engine systems engineers to stay under <40ms p99 retrieval budgets.
  • Build continuous distillation loops where heavy offline multi-document LLM synthesis judges generate billions of calibrated hard-negative relevance labels for real-time student rankers.
  • Pioneer query-rewriting, entity/temporal disambiguation, and dynamic sub-index routing models handling breaking news, code docs, financial filings, and scholarly literature.
  • Lead online interleaved experiments and end-to-end answer-quality evaluations impacting tens of millions of daily active users.

TECHNICAL ENVIRONMENT & OFFERING

Stack: Python, PyTorch, C++20/Rust, CUDA/TensorRT/ONNX, custom petabyte vector & inverted-index search engine, Ray, ClickHouse, multi-node H100/B200 GPU clusters. Workplace: Onsite in San Francisco, CA. Compensation: $255,000 – $355,000 base salary plus high-velocity unicorn equity, daily meals, and comprehensive US benefits.

Core Requirements

ESSENTIAL QUALIFICATIONS

  • 7+ years of machine learning engineering and applied research experience in web-scale search ranking, neural information retrieval, or large-scale recommendation/ads retrieval systems.
  • Deep track record training contrastive dense retrievers, SPLADE/learned sparse models, ColBERT-style late-interaction architectures, and distilled cross-encoder rerankers at multi-billion-document scale.
  • Strong systems awareness of ANN index trade-offs (RAM vs. NVMe, quantization recall loss, filtered search, real-time index updates) and sub-10ms GPU/SIMD model inference.
  • Mastery of counterfactual offline IR metrics, LLM-assisted relevance adjudication, and high-velocity online A/B experimentation.
  • Proven Staff-level ownership driving cross-team roadmaps across ML scientists and C++/Rust search systems engineers.

PREFERRED / BONUS EXPERIENCE

  • Prior Staff/Senior tenure on a major commercial web search engine, developer/enterprise search platform, or frontier RAG research team.
  • Publications at SIGIR, WSDM, WWW, ACL, EMNLP, or NeurIPS on neural retrieval, passage ranking, or retrieval-augmented reasoning.
  • Hands-on C++ or Rust familiarity for profiling inference-inside-index scoring kernels.

Position Telemetry

Location San Francisco, CA
Department Machine Learning
Employment type Full-time
Workplace type On-site
Salary range
$255k – $355k annual package