Enrico Piovano, PhD

Co-founder & CTO at Goji AI

Building the world's most advanced AI for personalized, conversational B2B demos that turn SaaS websites into high-converting sales engines

Ex-Amazon Applied Scientist (Alexa & AGI) · Expert in Agentic AI, LLMs & ML

Writing one of the most complete blogs on Agentic AI, LLMs, and real-world applications

PhD in Electrical Engineering, Imperial College London

Gold Medalist at the National Mathematical Olympiad

Technical Blog

About Me

Technical Expertise

I have built the entire Goji AI infrastructure from the ground up, integrating state-of-the-art Agentic AI systems and protocols (such as MCP and A2A), RAG systems, advanced search engine, LLM reasoning, fine-tuning, and reinforcement learning.

I have 6 years of experience in AI, including at Amazon, spanning Agentic AI, LLMs, and core ML including Transformer/BERT training and evaluation, offline-to-online performance correlation, data generation/cleaning, and production deployments.

Agentic AI
RAG Systems
Advanced Search
Engine
LLM Reasoning
ML Evaluation &
Production Correlation
Model Training &
Evaluation

Consulting & Coaching

Consulting

Expert guidance for organizations building production-grade AI systems that deliver measurable business impact.

  • Agentic AI & Multi-Agent Systems

    Architecture design, agent orchestration, protocol implementation (MCP, A2A)

  • AI Product Development

    RAG systems, prompt engineering, advanced search engines, chatbox

  • LLM Strategy & Implementation

    Fine-tuning, evaluation, and reinforcement learning

Coaching

Educational programs and knowledge transfer on Agentic AI, LLMs, RAG systems, and conversational AI to organizations upskill in the latest AI technologies.

  • Workshops & Seminars

    Hands-on AI implementation workshops and best practices for teams

  • Course Development

    Designing AI/ML training programs and educational curricula

  • Guest Lectures

    Delivering insights on AI applications to academic and professional audiences

Browse One of the Largest AI/ML Engineering Blogs

Deep dives from theory to applications

One of the most extensive AI/ML engineering resources available, covering state-of-the-art techniques with the most up-to-date knowledge.

Deep dives into architectures, production patterns, latest research papers, and hands-on implementation guides.

Experience

  • Multi-agent system: Connecting specialized agents and tools via MCP/A2A to seamlessly handle different tasks.
  • Advanced search engine: Create an advanced LLM-based search engine to retrieve the information in SaaS website in ChatGPT search style.
  • Deep context understanding: Using advanced tools and agents to extract provided context and perform background website searches to enrich the context.
  • Advanced Knowledge Retrieval: An advanced RAG system that surfaces the right information and visuals (like product tours) at the right time.
  • Conversational intelligence: LLM reasoning that provides the best answers and visuals, while proactively personalizing and engaging in the conversation.
  • Advanced Reasoning: Applying supervised fine-tuning (SFT) and reinforcement learning (RL) to enhance reasoning and improve performance over time.
Explore Goji
  • Worked on Agentic AI and LLMs training, fine-tuning, and evaluation.
  • Transformer-based natural language understanding (NLU), online performance prediction from offline metrics, synthetic data generation, dialog evaluation, offline prediction of dialog defects, and A/B testing.
  • Published 3 papers in international conferences.
  • Led several projects as team lead and managed two interns.

Selected achievements:

  • Optimized large-scale Agentic AI systems and LLMs for improved performance and efficiency.
  • Developed the key algorithm for predicting online performance from offline metrics, adopted globally across Alexa's NLU system.
  • Built and deployed an NLU deep neural network (DNN) model for the German locale — Alexa's first international deployment.
  • Designed and implemented new closed-loop beamforming algorithms for next-generation 802.11ah Wi-Fi modems.
  • Awarded the Roberto Padovani Scholarship (one of seven best-performing interns).

Research Intern

National Research Council, Turin, Italy

2012

Analyzed, designed, and optimized high-efficiency horn antennas.

Education

  • PhD thesis on the fundamental limits of cache-aided wireless networks under various types of uncertainties.
  • Investigated the high-SNR capacity of wireless networks with imperfect channel knowledge.
  • Designed deep learning–based wireless communication systems for simultaneous information and power transfer.
  • Published 3 papers in leading IEEE journals and 5 papers in top international conferences.
  • Supervised three MSc theses on MIMO networks with imperfect CSIT.

Double MSc in Electrical Engineering and Telecommunication Engineering

Telecom ParisTech, Paris, Eurecom, Sophia Antipolis, and Politecnico di Torino, Torino

2012 - 2015

Graduated with GPA 4.0 and 110 cum laude.

BSc in Telecommunication Engineering

Politecnico di Torino, Torino

2009 - 2012

Graduated with GPA 4.0 and 110 cum laude.

Awards

2009

Gold Medal

National Mathematical Olympiads

2008

Silver Medal

National Mathematical Olympiads

2013

Top Bachelor Graduate Student

Politecnico di Torino and Collegio Einaudi

Publications

Machine Learning & Conversational AI

  • Predicting Interaction Quality of Conversational Assistants With Spoken Language Understanding Model Confidences
    CIKM 2023, Applied Research Track
    Y. Gao*, E. Piovano*, T. Soliman*, et al. (*equal contribution)
  • Online Adaptive Metrics for Model Evaluation of Non-representative Offline Test Data
    ICPR 2022 (Oral Presentation - 10% acceptance)
    E. Piovano, D. T. Le, B. Chen and M. Bradford
  • It is better to Verify: Semi-Supervised Learning with a human in the loop for large-scale NLU models
    NAACL Workshop on Data Science with Human in the Loop, 2021
    V. Weber, E. Piovano and M. Bradford
  • A Learning Approach to Wireless Information and Power Transfer Signal and System Design
    IEEE ICASSP 2019
    M. Varasteh, E. Piovano and B. Clerckx

Information & Wireless Communication

  • Rate-splitting multiple access for overloaded cellular internet of things
    IEEE Trans. on Communications, 2021
    Y. Mao, E. Piovano and B. Clerckx
  • Centralized and Decentralized Cache-Aided Interference Management in Heterogeneous Parallel Channels
    IEEE Trans. on Communications, 2020
    E. Piovano, H. Joudeh and B. Clerckx
  • Generalized Degrees of Freedom of the Symmetric Cache-Aided MISO Broadcast Channel with Partial CSIT
    IEEE Trans. on Information Theory, 2019
    E. Piovano, H. Joudeh and B. Clerckx
  • Optimal DoF Region of the K-User MISO BC With Partial CSIT
    IEEE Communications Letters, 2017
    E. Piovano and B. Clerckx
  • Robust Cache-Aided Interference Management Under Full Transmitter Cooperation
    IEEE ISIT 2018
    E. Piovano, H. Joudeh and B. Clerckx
  • On coded caching in the overloaded MISO broadcast channel
    IEEE ISIT 2017
    E. Piovano, H. Joudeh and B. Clerckx
  • Overloaded multiuser MISO transmission with imperfect CSIT
    Asilomar Conference 2016
    E. Piovano, H. Joudeh and B. Clerckx

Contact

Open for consulting, coaching, and selective advisory opportunities. Whether you're building AI products or looking to upskill your team, I'd love to hear from you.