All jobs

Customer-facing LLM deployments are exactly her Halespring + Ledgerlight profile; Go and Postgres cover their integration stack.

Serving and eval work on open-weight models lines up with Signalcast and her vision-lab fine-tuning.

Building code-assistant features needs the prompt-eval loop she built in Signalcast; TypeScript depth checks out.

Signalcast is nearly this job: regression evals in CI, harness design, metric dashboards.

NL-to-SQL over customer data is close to Atlas Notes retrieval; missing their dbt layer.

Fraud-signal modeling maps to her two-tower ranking work; payments-domain experience from Brightpath is a direct hit.

Go services at Brightpath and the p99 latency story fit their storage-path team; distributed-systems coursework helps.

Eval-first team; Signalcast experience is the differentiator, Kubernetes less so.

Event-driven pricing needs the idempotent-API habits from Brightpath payments; freight domain is new.

Customer-facing instincts plus working code; Airflow certification not held.

Feature-store and pipeline work overlaps Halespring; their Ray stack is a gap.

Streaming ingestion matches her BigQuery feature-store build; no energy-sector background.

Eval-first team; Signalcast experience is the differentiator, Snowflake less so.

Streaming and Snowflake pipeline experience carries the data half of the role.

Retrieval work maps to Atlas Notes; their Kubernetes stack is a modest gap.

Solid stack match on Python and Snowflake; scope reads slightly senior.

Streaming and C++ pipeline experience carries the data half of the role.

Backend depth from Brightpath fits; Terraform exposure would need ramp-up.

Eval-first team; Signalcast experience is the differentiator, Rust less so.

Eval-first team; Signalcast experience is the differentiator, Rust less so.

Strong on APIs and Postgres; Terraform is listed as required and unproven.

Backend depth from Brightpath fits; Rust exposure would need ramp-up.

Streaming and C++ pipeline experience carries the data half of the role.

Fine-tuning record fits their model team; Kubernetes pipeline is new territory.

Recsys overlap from Halespring is strong; domain knowledge in Rust is thin.

Retrieval work maps to Atlas Notes; their C++ stack is a modest gap.

Fine-tuning record fits their model team; Snowflake pipeline is new territory.

Fine-tuning record fits their model team; dbt pipeline is new territory.

Vision-lab metrics work matches their dbt evaluation loop.

Eval-first team; Signalcast experience is the differentiator, Snowflake less so.

Customer-facing instincts plus working code; Airflow certification not held.

Strong on APIs and Postgres; Kubernetes is listed as required and unproven.

Solid stack match on Python and Snowflake; scope reads slightly senior.

Retrieval work maps to Atlas Notes; their dbt stack is a modest gap.

Eval-first team; Signalcast experience is the differentiator, Kafka less so.

Solid stack match on Python and Kubernetes; scope reads slightly senior.

Streaming and Kubernetes pipeline experience carries the data half of the role.

Vision-lab metrics work matches their C++ evaluation loop.

Solid stack match on Python and C++; scope reads slightly senior.

Recsys overlap from Halespring is strong; domain knowledge in Spark is thin.

Solid stack match on Python and Spark; scope reads slightly senior.

Fine-tuning record fits their model team; Rust pipeline is new territory.

Solid stack match on Python and Ray; scope reads slightly senior.

Fine-tuning record fits their model team; Snowflake pipeline is new territory.

Recsys overlap from Halespring is strong; domain knowledge in Snowflake is thin.

Vision-lab metrics work matches their Kubernetes evaluation loop.

Streaming and Rust pipeline experience carries the data half of the role.

Recsys overlap from Halespring is strong; domain knowledge in Spark is thin.

Fine-tuning record fits their model team; Spark pipeline is new territory.

Fine-tuning record fits their model team; Spark pipeline is new territory.

Vision-lab metrics work matches their C++ evaluation loop.

Backend depth from Brightpath fits; Kubernetes exposure would need ramp-up.

Retrieval work maps to Atlas Notes; their Kafka stack is a modest gap.

Customer-facing instincts plus working code; Kafka certification not held.

Backend depth from Brightpath fits; dbt exposure would need ramp-up.

Backend depth from Brightpath fits; Rust exposure would need ramp-up.

Recsys overlap from Halespring is strong; domain knowledge in Snowflake is thin.

Retrieval work maps to Atlas Notes; their Terraform stack is a modest gap.

Eval-first team; Signalcast experience is the differentiator, Kafka less so.

Eval-first team; Signalcast experience is the differentiator, Spark less so.