Staff/Senior Software Engineer, Hybrid IR and Systems

Helios; New York (NY); On-site; USD 120,000 - 180,000; Full time

Job summary

Helios is seeking a lead for the Helios Rapid Ontology System to own the end-to-end retrieval stack, including index design, hybrid retrieval, and reranking. You will advance query understanding, temporal retrieval, and ensure AI outputs are traceable and auditable.

Candidates should have deep expertise in information retrieval, vector and lexical indices, and modeling in regulated environments, and be a driven innovator capable of operating under high stakes.

Qualifications

• Requires deep expertise in retrieval mechanisms, vector and lexical indices, and temporal modeling.
• Candidates should be innovators capable of operating in high-stakes, regulated environments.

Responsibilities

• Own the end-to-end retrieval stack for the Helios Rapid Ontology System, including index design, hybrid retrieval, and reranking.
• Advance query-understanding and temporal-retrieval capabilities to ensure traceable and auditable AI outputs.

Skills

Information Retrieval

Vector Indices

Lexical Indices

Rank Fusion

Query Intent Classification

Entity Linking

Temporal Modeling

Multilingual Retrieval

Cross Encoder Reranking

BM25

HNSW

Tokenization

Federated Search

Embedding Models

Fine Tuning

Job description

Own the end-to-end retrieval stack for the Helios Rapid Ontology System, including index design, hybrid retrieval, and reranking. Advance query-understanding and temporal-retrieval capabilities to ensure traceable and auditable AI outputs.

Requirements: Requires deep expertise in retrieval mechanisms, vector and lexical indices, and temporal modeling. Candidates should be driven innovators capable of operating in high-stakes, regulated environments.

Key Skills: Information Retrieval, Vector Indices, Lexical Indices, Rank Fusion, Query Intent Classification, Entity Linking, Temporal Modeling, Multilingual Retrieval, Cross Encoder Reranking, BM25, HNSW, Tokenization, Federated Search, Capacity Planning, Embedding Models, Fine Tuning