I am a Senior AI Data Engineer focused on turning LLM-powered ideas into production features. I design agentic workflows, retrieval-augmented generation, and scalable AI services on AWS, with a bias for reliability, observability, and safe, privacy-preserving practices. I enjoy collaborating with product and platform teams to iterate from rapid prototyping to production monitoring, build guardrails, feedback loops, and regression suites, and help teams scale AI features responsibly.…

Thiago Assumpção

I am a Senior AI Data Engineer focused on turning LLM-powered ideas into production features. I design agentic workflows, retrieval-augmented generation, and scalable AI services on AWS, with a bias for reliability, observability, and safe, privacy-preserving practices. I enjoy collaborating with product and platform teams to iterate from rapid prototyping to production monitoring, build guardrails, feedback loops, and regression suites, and help teams scale AI features responsibly.…

Disponible para alquilar

I am a Senior AI Data Engineer focused on turning LLM-powered ideas into production features. I design agentic workflows, retrieval-augmented generation, and scalable AI services on AWS, with a bias for reliability, observability, and safe, privacy-preserving practices.

I enjoy collaborating with product and platform teams to iterate from rapid prototyping to production monitoring, build guardrails, feedback loops, and regression suites, and help teams scale AI features responsibly.

Ver más

Nivel de experiencia

Experto
Experto
Experto
Experto
Experto
Experto
Experto

Idioma

Inglés
Fluido

Experiencia laboral

Senior AI Data Engineer at Choreograph (WPP)
August 1, 2024 - November 18, 2025
Led LLM features in production with an AWS-first (and GCP-hybrid) setup. Designed a LangGraph-based agent orchestrating tools for audience lookup, budget planning, and channel mix, persisting memory/state in PostgreSQL for multi-step sessions. Integrated the agent into the existing FastAPI backend with modular routers, DI, OIDC auth, and quota controls; documented via OpenAPI. Instrumented LangSmith for trace-level monitoring/evaluation and built dashboards for latency, tool errors, and drift signals across prompts and retrieval strategies. Shipped document intelligence & extraction via RAG (LangChain) over briefs and historical performance docs; curated prompt templates tuned to marketing use cases. Established feedback loops (annotation queues, error triage) and introduced prompt/version stores and regression suites. Partnered with platform/data teams to improve observability and scale on EKS/Lambda; IaC + CI/CD with Terraform/GitHub Actions.
Senior Data Engineer at Etsy (Depop)
July 1, 2024 - July 1, 2024
Orchestrated Spark pipelines on Kubernetes with Airflow to supply features/datasets for ML & analytics; improved text-data quality gates. Authored internal Python SDKs to standardise dataset access and feature computation; improved developer productivity and consistency. Integrated dbt models and data lineage (DataHub); documented contracts and exposures for downstream consumers. Automated infra with Terraform; added CI checks and deployment pipelines (GitHub Actions) to increase reliability.
Senior Data Engineer at BlueOrange
1 de enero de 2022 - 1 de enero de 2022
Introduced Kafka AVRO + Schema Registry patterns; standardised ingestion and schema evolution for downstream analytics. Built Snowflake and BigQuery models; implemented cost/perf practices (partitioning, clustering, task scheduling). Containerised Python services (Docker), established CI/CD, and delivered reliable data products for analytics teams.
Senior Software Engineer at BJSS
May 1, 2020 - May 1, 2020
Built secure APIs and automated CI/CD for government platforms; contributed to observability and performance tuning. Worked across teams to integrate services with reporting/analytics layers and testing frameworks.
Senior Backend/Data Engineer at Telecine
December 31, 2019 - December 31, 2019
Built high-throughput Flask microservices backed by DynamoDB/DocumentDB with retries and idempotency. Delivered event pipelines (Kinesis Firehose → S3 → Glue → Redshift) with cataloguing and partitioning for downstream analytics.

Educación

Electrical Engineering at São Paulo State University
January 1, 2003 - January 1, 2009

Requisitos

Añade aquí tus titulaciones o premios.

Experiencia en el sector

Software & Internet, Professional Services, Media & Entertainment