Thiago M. Barros

AI Engineering Lead · Building AI products from zero to launch

AI/ML lead and founding engineer with 7+ years of experience building AI products across industrial ventures, consulting, and life sciences. At a Nokia Bell Labs venture, built the product's server-side platform and LLM agent from zero as the platform's only engineer, took it to launch with 3 mining customers, and now lead its 8-engineer team and product roadmap.

Highlights

Experience

Nokia São Paulo, Brazil

AI/ML Lead · Cognitive OperationsMar 2025 to present

  • Joined a Nokia Bell Labs venture at the idea stage as AI lead and founding engineer of its server-side platform, building it alone for the first year and taking it to launch as Nokia Cognitive Operations (June 2026), a real-time operational picture for mining, emergency services, and defense, with 3 mining customers
  • Built an enterprise LLM agent (RAG, OpenAI Responses API, 23 custom tools) for natural-language data querying and in-app actions, serving 2,480 production queries and tested nightly against 43 scenarios
  • Owned the platform's product roadmap, prioritizing features from customer discovery, site visits, and sales input, and replaced a third-party digital twin with an in-house solution that increased margin by 9% (pilot)
  • Grew from sole platform engineer to leading an 8-engineer team that ships quarterly releases, translating the roadmap into technical specs and architecture while still building core features
  • Delivered the platform end to end, writing over half of its 250K lines of code: 28+ microservices, real-time dashboards, geospatial maps, alerting, and Azure AD SSO, deployed with Helm on Azure AKS, on-prem VMware, and k3s
  • Managed the Microsoft partnership that brought the product to Azure Marketplace
  • Designed the platform's test strategy to pass Nokia's release quality gates, leading the team to 2,900+ automated tests run in an 18-stage CI/CD pipeline
  • Led development of embedded AI on Hailo and NVIDIA Jetson for fatigue detection, anomaly detection, and LiDAR compression, now in customer pilots, cutting uploads 49x in edge-to-cloud tests (up to 192x per frame)

Accenture São Paulo, Brazil

AI Decision Science Consultant · Applied IntelligenceFeb 2023 to Mar 2025

  • Led client teams of 2 to 6 consultants across 3 projects, as technical lead or project manager, translating customer pain points into custom applied intelligence products
  • Developed a simulation web app prescribing inventory policies that generated R$ 88M in savings
  • Reduced forecast error by over 10% in 6 months with a custom regression model, adding R$ 50M in annual revenue
  • Implemented an NLP neural network to extract features from over 10M part descriptions, improving marketplace pricing accuracy by 31%
  • Implemented a custom supply chain simulator algorithm that increased revenue by R$ 2M monthly

AbbVie Chicago, IL, USA

Chemist II · Analytical R&DNov 2019 to Jul 2022

  • Developed ML models of molecular degradation to support drug shelf-life submissions to the FDA and other regulators
  • Designed and coordinated 200+ studies and the release of 400+ GMP batches for clinical use across 80+ projects

University of Illinois research labs Champaign, IL, USA

Research Assistant2016 to 2018

  • Collected and analyzed neural and behavioral data from 80+ subjects in a neuropsychopharmacology lab

Skills

AI/ML
LLM agents (RAG, tool calling, evals), edge AI and computer vision (Hailo, Jetson), NLP, deep learning, forecasting, simulation, MLOps
Engineering
Python (PySpark, pandas, TensorFlow), SQL, Kubernetes (Helm), Azure (AKS, Bicep), AWS, Databricks, MQTT, InfluxDB, Keycloak, GitLab CI
Leadership & product
Team leadership, product roadmaps and specs, customer discovery, build-vs-buy decisions, enterprise partnerships, security reviews
Languages
Portuguese and English (bilingual)

Education & certificates

University of Illinois at Urbana-Champaign, B.S. in Chemistry and Psychology (Behavioral Neuroscience), 2019. Dean's List.