- Company Name
- Lorven Technologies Inc.
- Job Title
- Senior Engineer (Agentic + Data/ML)
- Job Description
-
**Job Title**
Senior Engineer – Agentic Retrieval, Data & ML
**Role Summary**
Design, implement, and maintain agentic retrieval workflows, scalable data pipelines, and ML/LLM‑based scoring systems that drive engineering productivity and code‑quality metrics for a large video‑content provider’s distributed software teams.
**Expectations**
- 5–10+ years as a software, data, or ML engineer.
- Advanced Python programming with modern data/ML tooling.
- Proven experience delivering production‑grade orchestration (Dagster or Airflow) on AWS.
- Hands‑on expertise in LLM prompting/optimization (DSPy preferred).
- Strong knowledge of Git/PR metadata, code diffs, backfills, and performance‑critical data pipelines.
**Key Responsibilities**
- Build and optimize CES (Code‑Evaluation‑Score) improvements: PR‑level scoring, historical context extraction, automated score correction from feedback.
- Ingest and link PRs, commits, branches, and merge events; manage backfills and data lineage.
- Migrate, schedule, and monitor DAGs with Dagster; implement retries, observability, and data‑quality alerts.
- Design and tune DSPy‑based prompts, evaluation sets, and cost/quality trade‑offs for LLM inference.
- Develop backend services to ingest metadata, compute quality metrics, and expose them via API/UI.
- Implement classification of work types, epic linking, and filtering/explorer features.
- Run evaluation loops, gather user feedback, and close the loop via automated or manual score corrections.
**Required Skills**
- **Agentic Coding**: Retrieval agents that pull code context, diffs, and history.
- **Data Orchestration**: Dagster (preferred) or equivalent; AWS services (ECS, EventBridge, Lambda, S3, Glue).
- **LLM & Prompting**: DSPy or similar, prompt chaining, context packing, evaluation design.
- **Data Modeling & Analytics**: PR/commit linking, branch/main tracking, anomaly detection.
- **Backend Integration**: Build services for metadata ingestion, score computation, and API exposure.
- **Feedback & Evaluation**: Design eval sets, collect feedback, perform automated/manual corrections.
**Nice to Have**
- Experience with code‑scoring or effort/quality metrics and SonarQube integration.
- Graph‑like linking across artifacts (Jira, PR, commit, branch).
- LLM performance/cost tuning (batching, caching).
- Observability for pipelines (metrics, traces, logs) and data‑quality alerting.
**Required Education & Certifications**
- Bachelor’s or Master’s degree in Computer Science, Engineering, Data Science, or related field.
- No mandatory certifications, but familiarity with AWS Certified Developer or similar is advantageous.