Siavosh Mahboubian
Mentor
(22 reviews)
US$24.00
For every 15 mins
28
Sessions/Jobs
ABOUT ME
Senior Data Scientist: LLMs, RAG and graph ML. MSc in ML from UCL.
Senior Data Scientist: LLMs, RAG and graph ML. MSc in ML from UCL.
I'm a senior data scientist with nine years of experience taking ML systems from idea to production, specialising in Large Language Models, RAG and graph analytics.
I recently completed a contract at the UK's Financial Conduct Authority, building graph analytics on AWS Neptune to help uncover hidden financial-crime networks. Before that, at Phaidon International, I designed and deployed Bullseye, an AI candidate-matching system (Neo4j graph of 2M+ candidates, OpenAI embeddings, LLM reranking) that became the firm's primary recruitment technology, used by 100+ recruiters across three countries.
I hold an MSc in Computational Statistics and Machine Learning from UCL, with thesis research on link prediction in criminal networks, first applied during a placement with the UK Home Office's organised crime division.
I enjoy making complex things easy to understand. Whether you need help debugging Python, designing an ML pipeline, integrating LLMs or RAG into your product, or getting a graph database to behave, I can help.
English
London (+01:00)
Joined July 2019
EXPERTISE
9 years experience | 10 endorsements
MSc in Computational Statistics and Machine Learning (UCL). Nine years of production ML: predictive modelling, anomaly detection, model v...
MSc in Computational Statistics and Machine Learning (UCL). Nine years of production ML: predictive modelling, anomaly detection, model validation, recommender systems.
12 years experience | 5 endorsements
Primary language for over a decade: FastAPI backends, ML pipelines (scikit-learn, PySpark), LLM integrations.
Primary language for over a decade: FastAPI backends, ML pipelines (scikit-learn, PySpark), LLM integrations.
9 years experience
SQL and graph databases in production: MySQL, PostgreSQL, Neo4j, AWS Neptune.
SQL and graph databases in production: MySQL, PostgreSQL, Neo4j, AWS Neptune.
8 years experience | 2 endorsements
9 years experience
3 years experience
Production LLM systems since 2023: RAG architectures, embeddings, prompt engineering and evaluation (DSPy), OpenAI and Anthropic APIs.
Production LLM systems since 2023: RAG architectures, embeddings, prompt engineering and evaluation (DSPy), OpenAI and Anthropic APIs.
9 years experience
Link prediction, community detection, entity resolution (Splink). Production graph systems from criminal networks at the Home Office to f...
Link prediction, community detection, entity resolution (Splink). Production graph systems from criminal networks at the Home Office to financial-crime networks at the FCA.
REVIEWS FROM CLIENTS
(22 reviews)
Parth Agrawal
November 2019
Sio is great! He spends ample time with his prep work before each session so that he can be prepared and provide the best, most optimal tutoring session possible. He really cares about his students and has a genuine interest to help students succeed and understand material.
Parth Agrawal
October 2019
Thanks Sio! Sio is really good at explaining material and is great at making difficult problems look easy.
Parth Agrawal
October 2019
Excellent as always. Really helped me understand a difficult mathematical optimization problem for Machine Learning.
Parth Agrawal
October 2019
Sio always goes above and beyond. He has a great grasp on machine learning topics and always does preparation before each one of our sessions providing the most value for the time
Zachary Kimpfler
October 2019
Like always, Sio is friendly, patient, knowledgeable, and accommodating when working together. I'd highly recommend him!
Parth Agrawal
October 2019
Sio's lessons are extremely helpful. I always come away with new knowledge and get answers to material I am unsure about.
Parth Agrawal
October 2019
Excellent as always. Sio really takes the time to go through and explain material thoroughly so that one can have fundamental and foundational knowledge to master the advanced topics in the subject area of machine learning.
Zachary Kimpfler
September 2019
I've worked with Sio several times now and he goes above and beyond my expectations every time! He is very knowledgeable, patient, and he always make sure I comprehend the information we are talking about before proceeding with our lesson. I can't recommend him enough to anyone having trouble with machine learning topics or any mathematical topics along those lines! Awesome!
Zachary Kimpfler
September 2019
Sio is an awesome mentor! He is always quick to respond, very patient with any and all questions, and very knowledgeable. He always makes sure I comprehend the material at hand and goes the extra mile to make sure that everything is correct and clearly explained! Again, I cant recommend him enough!
Khandakar Rafid Rahman
September 2019
Sio is very smart and cooperative. He is very helpful and can clarify any doubts. I especially appreciate him explaining each step, for better clarification. His knowledge in Data cleansing, really showcased when he helped me out with a project. I highly recommend him!
SOCIAL PRESENCE
75 Reputation
1
1
5
EMPLOYMENTS
Graph Data Science Specialist (Contract)
Financial Conduct Authority (FCA)
2026-04-01-2026-07-01
Built graph-analytics tooling on the FCA's AWS Neptune network platform to surface hidden financial-crime networks. Designed a 39-check a...
Built graph-analytics tooling on the FCA's AWS Neptune network platform to surface hidden financial-crime networks. Designed a 39-check assurance framework for CIAM, the FCA's ML system for ranking firms by financial-crime risk.
Python
Machine Learning
Graph Database
View more
Python
Machine Learning
Graph Database
Amazon Neptune
View more
Senior Data Scientist
Phaidon International
2025-03-01-2026-02-01
Designed and deployed Bullseye, an AI-powered candidate matching system (Neo4j graph of 2M+ candidates, OpenAI embeddings, LLM reranking,...
Designed and deployed Bullseye, an AI-powered candidate matching system (Neo4j graph of 2M+ candidates, OpenAI embeddings, LLM reranking, FastAPI) that became the company's primary recruitment technology across UK, US and German offices, adopted by 100+ recruiters. 2.7x improvement in placement rates vs baseline in backtesting.
Python
Neo4j
OpenAI
View more
Python
Neo4j
OpenAI
Fastapi
LLM
View more
Senior Data Scientist
ProFinda
2023-02-01-2025-02-01
Built an LLM-powered skill ontology engine combining Neo4j knowledge graphs, vector embeddings and LLM inference. Extended RAG architectu...
Built an LLM-powered skill ontology engine combining Neo4j knowledge graphs, vector embeddings and LLM inference. Extended RAG architectures to graph-structured data. Designed prompt evaluation pipelines with DSPy. Work supported an HMRC R&D innovation grant.
Python
Neo4j
LLM
View more
Python
Neo4j
LLM
RAG
View more
PROJECTS
Link Prediction in Organised Crime Networks
UK Metropolitan Police, UK Home Office
2017
Project was done as part of my masters thesis with the UK Home Office and, more specifically, with the Organised Crime Division of the Me...
Project was done as part of my masters thesis with the UK Home Office and, more specifically, with the Organised Crime Division of the Metropolitan Police. Aim of the project was to assess the reliability of various Link Prediction Algorithms on the classified datasets held by the UK government. This project required me to obtain a UK National Security Clearance prior to starting.
Python
Neo4j
GraphQL
View more
Python
Neo4j
GraphQL
View more
MSc Computational Statistics and Machine Learning
2017
Python
Machine Learning
Statistics
Python
Machine Learning
Statistics