Alessandro Pedori
Verified Expert in Engineering
Artificial Intelligence Developer
Berlin, Germany
Toptal member since November 18, 2022
Alessandro is a full-stack AI, NLP, and ML engineer, consultant, and architect with several years of experience teaching AIs what to do and building the code around it, mainly in Python. He also loves helping people understand the limits of current systems, triaging what to implement first, building code, teaching, consulting, mentoring, and coaching. Above all, Alessandro is passionate about how technology can empower humans to shine.
Portfolio
Experience
Availability
Preferred Environment
MacOS, Linux, Python, Visual Studio Code (VS Code)
The most amazing...
...tool I've developed is a system to understand the intention of not renewing a subscription from client emails hinting to churn.
Work Experience
Co-founder | CTO | Coach
IFS Collective
- Co-created the offering of the company, shaping the goals and online presence.
- Envisioned, prepared, and facilitated online unconferences, adapting the software and tools to the online setting.
- Ideated, developed, and tested apps using AI to support IFS and other talk therapies.
Industry Leader | AI Business Consultant Course
Hyper Island
- Developed a curriculum that helps beginners get hands-on experience in programming and performing preliminary actions.
- Found, briefed, and coordinated workshop facilitators hosting lessons for the students.
- Triaged student ideas and developed projects with them.
Chief Technology Officer
EnterpriseAI Consulting UG
- Designed proof of concept and minimum viable products for several consulting projects, including requirements and implementation.
- Performed feasibility studies, triaged the required skills, set up and conducted necessary interviews, and performed high-level project management of the teams.
- Hosted workshops on data maturity and AI readiness for small and medium enterprises.
Senior P2P Interviewer for Data Science and Python Verticals
Job Board for Freelance Work
- Structured interviews for experienced Python programmers and created the curriculum for the data science vertical, including data scientists, data engineers, machine learning engineers, and data analysts.
- Interviewed 100+ candidates, from middle to CTO level, evaluating them on problem-solving, technical skills, and live coding.
- Developed proposals for live coding, allowing non-exerts to participate in technical interviews.
MVP Developer
Hearst - Technology
- Developed an app for Hearst media group to create full car listings from photos and replies to questions, using previous listings as models.
- Built an app to extract meaningful information from a list of devices to group them, using primarily LLMs, GPT, and heuristics.
- Developed an app to automate website scraping using a mixture of heuristic, NLP, and OpenAI APIs.
- Created an app to categorize car photos using zero shot pre-trained transformer models.
- Developed an app that scrapes and catalogs car data to create replies to questions to create car listings using OpenAI APIs.
Speaker Identification Model Consultant
GovWorx, Inc.
- Developed and fine-tuned language models to detect actions from transcriptions of interactions.
- Researched state-of-the-art developments for speaker identification, intent attribution, and summarization of long transcripts, then proposed implementations.
- Explored using available datasets and synthetic data for pre-launch development.
Lead Research ML Engineer
molab.ai
- Developed and compared transformer-based models to predict molecule absorption, distribution, metabolism, excretion, and toxicity (ADMET) features from the simplified molecular-input line-entry system (SMILES) representation.
- Adapted published research code to the production environment.
- Created several data pipelines to test different models for ADMET prediction.
Senior ML Engineer
Ella Lab
- Took the models and software from proof of concept to deployment, having them react in the time required.
- Explored all possible ways to deploy huge transformer-based language models on available resources, working with bleeding edge architectures and constantly changing models.
- Oversaw the optimization and parallelization of the architecture, which resulted in a 10x reduction in the processing time.
- Deployed and tested big language model-based APIs in production, managing GPU out or memory errors and a continuously changing NLP system.
Chief Data Scientist
Staircase
- Implemented ways to extract behavior from customer emails, differentiating general emotions from the intention not to renew contracts or services.
- Developed ways to extract company information from emails, calendar invites, and meeting transcriptions. The extracted data included the role of different people, their capacity to influence decisions, and changes in their roles.
- Oversaw design and re-implementation of several services, including traditional NLP, regular expressions, transformers, and deep learning.
- Supported the transition to using and deploying transformers for NLP in production.
- Consulted on appropriate technological solutions and feasibility.
Lead ML Engineer
Klarna
- Developed MLOps systems with very low latence to maintain high availability under high loads.
- Evaluated several platforms to enable data scientists to deploy models without involving engineers (hint: not recommended).
- Created several proofs of concept for ML and DS process improvements.
Senior ML Engineer
Miele
- Designed and built microservices architecture serving different ML models for shape recognition, categorization, text detection with subsequent OCR, etc.
- Set up DevOps architecture for other developers and remote access.
- Oversaw the curation of image data, the selection of text detection, optimization and recognition algorithms, and object recognition.
Senior NLP Engineer
Retresco
- Extracted and organized gigabytes of data from Wikidata, restructuring the process to parallelize and speed up from more than two days to less than four hours.
- Implemented NLP processes to disambiguate pieces of information from Wikidata, creating an IBM Watson-like system to answer general knowledge queries.
- Restructured the architecture of a complex system already in deployment, allowing a better evolution of the platform and bug fixing.
Research Engineer in NLP and ML
Aaron
- Applied NLP and ML techniques to automate customer support tasks.
- Performed data analysis, clean-up, normalization, and topic modeling on business and customer support data.
- Designed and implemented Alexa skills-based voice assistants for financial services, including checking account balances and blocking and requesting credit cards.
Senior Python Back-end Developer
UberResearch
- Developed ways to parallelize the code using ZeroMQ, cutting the execution time from days to hours.
- Oversaw the code development required to ingest several gigabytes of data in the system, keeping it up to date on AWS without replications.
- Planned, designed, implemented, and refined ETL pipelines for the data processes.
Support and Sales Engineer
Acrolinx
- Developed an app to collect user feedback on the support team via simple links in emails (a the time, it was new) and visualized it on a dynamic page on the company website.
- Tracked and fixed bugs, supporting customers in adapting the software provided to their needs.
- Expanded the language capabilities of the software to support the different needs of clients.
Senior Localization Engineer for Siri
Apple
- Implemented ways to adapt Siri's expressions to Italian for its first international release.
- Kept track of bug priorities and coordinated with translators, testers, and other engineers.
- Oversaw the triaging and fixing of important edge cases, including pre-empting ways in which users would try to get Siri to be vulgar or inappropriate.
Localization Engineer
Snap-On
- Built a tool to automate daily localization checks, going from a two-hour daily workload to a five-hour one.
- Downloaded, compared, and checked 16+ million words of localized content regularly, coordinating with translators and engineers.
- Developed a system to compare inventory items between departments, scraping inventory websites and connecting databases.
Experience
General Intent Detection from Customers' Emails
https://staircase.ai/I was the main back-end developer. I also developed the algorithms based on NLI transformer models and some additional custom rules. The customer was so happy with the system they decided to apply to patent the technique; if accepted, my name would be on the patent.
Voice-activated Shopping Based on Alexa
I was the main designer, back-end developer, NLP developer, and interaction designer. This tool was created in 2018, when the ideal use of voice-activated assistants was unclear. The project was a tremendous technical success and proved that users did not want to shop for groceries talking to Alexa.
Telegram Voice Messages Transcription Bot
Gender B(l)ender
Adding German Support to LangCheck
https://github.com/ischender/langcheckEducation
Master's Degree in Computer Engineering
University of Bologna - Bologna, Italy
Skills
Libraries/APIs
PyTorch, Scikit-learn, Social Media APIs, Natural Language Toolkit (NLTK), ZeroMQ, Keras, OpenCV, TensorFlow, Telegram Messenger API, Pandas, SpaCy, Google Speech API
Tools
ChatGPT, Git, Google AI Platform, Google Bard, Microsoft Copilot, Open Neural Network Exchange (ONNX), Amazon SageMaker, Amazon Textract, NLPP, Dialogflow, Amazon Lex, AWS Step Functions, Whisper
Languages
Python, Elixir, SQL, Python 3, Regex, JavaScript, Perl, Java, HTML, PHP, XML
Paradigms
Agile, Rapid Prototyping, Rapid Application Development (RAD), Automation, Test-driven Development (TDD), ETL, DevOps, Parallel Programming, Agile Project Management
Platforms
MacOS, Visual Studio Code (VS Code), Unix, Linux, Amazon Web Services (AWS), Amazon, Amazon Alexa, Docker, NVIDIA CUDA, Google Cloud Platform (GCP), Ubuntu, Windows, Heroku, Azure, Firebase
Frameworks
Flask, Pyramid, Bootstrap, Laravel, gRPC, Svelte, Bedrock, LlamaIndex
Storage
Data Pipelines, Elasticsearch, PostgreSQL, Google Cloud, Google Cloud Datastore, Amazon S3 (AWS S3)
Industry Expertise
Healthcare, Social Media, Retail & Wholesale, Insurance
Other
Programming, Artificial Intelligence (AI), Natural Language Processing (NLP), Consulting, Technical Consulting, Workshop Facilitation, Machine Learning, Chatbots, Workshops, APIs, Rapid Development, Communication, Text Generation, Language Models, AI Design, Early-stage Startups, Startups, Coaching, Generative Pre-trained Transformers (GPT), Generative Pre-trained Transformer 2 (GPT-2), Text Processing, Software Architecture, Technical Leadership, NLU, Document Parsing, Deep Neural Networks (DNNs), BERT, Conversational Interfaces, Text Recognition, Large Language Models (LLMs), Data Processing, Data Transformation, CSV, OpenAI GPT-3 API, Artificial General Intelligence (AGI), Hugging Face, Fine-tuning, Minimum Viable Product (MVP), GPT-3, API Integration, Writing & Editing, Model Development, Algorithms, CSV File Processing, AI Chatbots, Custom Models, Generative Artificial Intelligence (GenAI), Text-to-text Transfer Transformer (T5), Data Scraping, PDF Scraping, Web Scraping, Training, Sentiment Analysis, Generative Pre-trained Transformer 4 (GPT-4), Artificial Neural Networks (ANN), Machine Learning Algorithms, System Design, Startup Consulting, Public Speaking, Interviewing, Deep Learning, OCR, Speech Recognition, Architecture, Leadership, Open Source, Health, Voice Analysis, OpenAI, Data Science, Cloud Platforms, Full-stack, PDF, Data Analytics, Data Analysis, Text to Speech (TTS), Frameworks, Speech to Text, OpenAI GPT-4 API, Voice Recognition, Speech Synthesis, Content Writing, Data Matching, Topic Modeling, Data Engineering, Llama 3, LangChain, Speech to Text AI, CI/CD Pipelines, Copywriting, Large Language Model Operations (LLMOps), Networking, Siri, Statistics, Team Management, Containerization, User Experience (UX), Voice, Distributed Software, Machine Learning Operations (MLOps), University Teaching, Facilitation, Wikidata, Tornado, Computer Vision, Telegram Bots, Image Processing, Web Development, Plugins, Google BigQuery, Healthcare IT, Psychology & Mental Health, DaVinci, Localization, Localization Testing, IT Support, Email Support, Google Cloud ML, Data Reporting, Graphics Processing Unit (GPU), Automatic Speech Recognition (ASR), Speaker Identification (SI), Prompt Engineering, Retrieval-augmented Generation (RAG), Chatbot Conversation Design, eCommerce, Inventory, Legal Technology (Legaltech), Data Visualization, Regression Modeling, Statistical Modeling, LoRa, Electronic Health Records (EHR), Graph Neural Networks
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring