Sergio Verga

AI Engineer | ML Engineer

Sergio VergaBuilding reliable data and AI pipelines for LLMs, robotics, and analytics.

AI/ML engineer with an MSc in Data Science and a BSc in Physics. I design data pipelines, evaluation loops, and LLM/RAG workflows that move research prototypes into reliable systems.

Download CVContact
FocusLLM/RAG
ResearchAgents/Robotics
SystemsEvaluation

//Featured Projects

A focused snapshot of recent work spanning robotics, computer vision, and GenAI pipelines.

View all projects
Automatic KPI Interpretation with Multimodal LLMs

Automatic KPI Interpretation with Multimodal LLMs

Interpreting business KPI dashboards with multimodal LLMs inside KNIME.

  • Built a modular LLM workflow to explain KPI shifts with structured summaries.
  • Highlighted anomalies and trends using multimodal dashboard inputs.
  • Delivered consistent, human-readable explanations for stakeholders.
GenAIKNIMELLMAnalytics
KNIME Hub
Vision-Language-Action (VLA) Models for Robotics

Vision-Language-Action (VLA) Models for Robotics

Multimodal agents for robotic pick-and-place following prompt-specified features.

  • Research activity (band 25CE114) with currently private repositories (paper under submission).
  • Benchmarked teleoperation interfaces (with 1-to-1 kynematic map) to train Imitation Learning models.
  • Experimented extensions of ACT model to support multi-tasking.
RoboticsVLAMultimodalDatasetEvaluation
LeRobotLeRobot Docs

//Experience

Research and engineering roles with a focus on data pipelines, evaluation, and ML systems.

Full experience
Intelligent Sensing Lab (ISLab), Univ. of Milano-Bicocca logo

External Research Collaborator (AI for Robotics)

Intelligent Sensing Lab (ISLab), Univ. of Milano-Bicocca
Mar 2025 - Oct 2025
  • Research project: “Data Collection and Analysis of Human–Machine Interaction in the Context of Industry 5.0” (Cod. 25CE114).
  • Part of HOMEY – A Human-centric IoE-based Framework for Supporting the Transition Towards Industry 5.0 (PRIN 2022, ID 2022-NAZ-0329/PER, funded by the EU – NextGenerationEU).
  • Collected and analyzed human-robot interaction data using the LeRobot framework.
  • Extended ACT policy models with visual task encoding, evaluated multi-task learning performance, and benchmarked wearable-sensor-based control (Movella Xsens Dot) vs. leader-arm teleoperation.

//Writing

Short technical pieces on analytics, KPI interpretation, and ML applications.

Read all

//Contact

Open to data engineering and AI engineering roles, collaborations, and research projects.

Email SergioLinkedIn