Bruno Volpato

Datadog AI Research: Contribute to retrieval paths, synthetic episode generation, ground-truth labels, and evaluation loops for observability agents. Helped kickstart an operational investigation agent, presented it at Engineering all-hands, and helped secure dedicated product and engineering staffing.

Simulation environments: Generate operational environments with controlled disruptions, telemetry, and ground-truth investigation paths for agent evaluation and RL training. Helped grow the catalog from roughly a dozen hand-built applications to thousands of generated applications.

Query federation: Designed and shipped Connector Server to decouple data-source integration from the query engine. Expanded federation from 5 to 40+ distinct sources for product and AI Research teams. The connector platform cut new-source onboarding from quarters to days.

Query platform: Previously led query planning and connectors serving hundreds of millions of queries daily. Delivered 50x speedups for customer-facing Logs Analytics production queries and reduced a large-timeframe query workload from ~34 seconds to 2.4 seconds (14x).

Leadership & reliability: Formed Query Connectors and grew it to 12 engineers as a shared platform team; mentored across levels and conducted 50+ design and coding interviews. Serve as a Core Incident Commander for severe incidents: 30+ commanded and 120+ responses.

Open query systems: Became a Substrait committer (Cross-Language Serialization for Relational Algebra) and contributed across DataFusion, Trino, Iceberg, and Calcite so plans and data access are portable across engines.

Data platform leadership: Led Dataflow Platform team of 5 engineers responsible for 200+ production pipelines processing petabytes of data daily across Google Cloud.

Reusable pipelines: Moved the team from bespoke customer pipeline delivery toward reusable infrastructure for teams and partners; reduced new pipeline development time by 60% and enabled 3x more use cases with the same team capacity.

Operational reliability: Participated in 24/7 on-call handling P0/P1 incidents, customer escalations, and turnarounds while improving tests, probers, rollouts, support automation, and release safety.

Container security: Led migration to minimal Distroless container images, eliminating over 80% of CVE surface area, and drove multi-quarter migrations and partner launches across data platform teams.

Streaming infrastructure: Architected scalable streaming pipelines processing billions of events daily using Kafka, Elasticsearch, and Couchbase, powering ML-driven recommendation engines for enterprise customers.

Rule execution: Built a dynamic configuration-to-bytecode compilation engine, reducing rule evaluation latency by 80% and enabling real-time personalization at scale.

Enterprise integration: Designed and implemented a data integration layer serving 5,000+ global manufacturing companies, processing millions of transactions daily with sub-second latency requirements.

Developer productivity: Built source search and white-box validation tooling, improving team productivity by 30% and reducing bug escape rate.