APPLIED ML RESEARCHER & TECHNICAL LEADER

Spyros
Mouselinos.

Applied ML researcher and technical leader. I build multi-agent systems at eBay. Previously, I worked on generative models at Moonvalley and production ML at DeepSea.

Explore my work Download CV

Amsterdam, NL / Currently building at eBay

Short on time? Start with the essentials.

Fleet MLAnomalies, sparse data & routing
DeepSea Technologies

Diffusion modelsDistributed video-model training
Moonvalley

Agents → actionRecommendations & automation
eBay Ads

ICLR · ACL · EMNLPFirst-author research
Reasoning & robustness

01 / SELECTED WORK

Research instincts.
Real-world outcomes.

Different domains. The same thread:
understand the problem, build the system,
make it count.

AGENTIC AI01

eBay / 2026–Present

From signals to decisions

Coordinated agents that discover advertising opportunities, make recommendations, and automate internal workflows.

Multi-agent systemsApplied research

Inside the work

The challenge

Turn ambiguous Ads research problems into useful recommendations and automated technical workflows.

My contribution

Architected agent roles, orchestration, and implementation from the ground up, owning the path from problem formulation to working systems.

The outcome

Delivered agentic systems across distinct eBay Ads use cases, connecting discovery, reasoning, and action.

Read the eBay work story ↗

MULTIMODAL GENERATIVE AI02

Moonvalley / 2024–2026

Training generative models at scale

Large-scale distributed model training, alongside research on visual quality and temporal consistency.

Distributed trainingTensor & data parallelism

Inside the work

The challenge

Train large generative models across distributed compute while maintaining visual quality, temporal consistency, and data provenance.

My contribution

Built parallel training code with model sharding and tensor and data parallelism. Also led R&D on rendered-text fidelity and motion-aware placement, and developed preference-optimization workflows for 70B-parameter LLM/VLM assistants.

The outcome

Trained large models using distributed infrastructure and developed synthetic-data pipelines spanning hundreds of millions of licensed video–text pairs.

Read the Moonvalley work story ↗

PRODUCTION ML03

DeepSea Technologies / 2018–2021

Making sense of imperfect vessel data

Anomaly detection, missing-data modeling, and weather-aware vessel routing across sparse, irregular, and corrupted time series.

Anomaly detectionSparse time seriesWeather-aware routing

Inside the work

The challenge

Extract reliable operational signals from missing observations, inconsistent sampling, corrupted sensor data, and changing weather conditions.

My contribution

Led a five-engineer ML platform team building anomaly detection, vessel-performance models, predictive maintenance, and weather-aware routing. Worked on missing-data imputation and productionized a multimodal Transformer with monitoring and over-the-air updates.

The outcome

20M predictions daily; 50% less bunker-fuel fraud across 1,000 vessels; routing for 100+ vessels. Additional models reduced fuel consumption by 4% and maintenance costs by 10%.

Read the DeepSea Technologies work story ↗

REAL-TIME ML SYSTEMS04

Equinox AI B.V. / Jun–Sep 2024

From market data to controlled execution

Low-latency inference, reinforcement-learning agents, and distributed backtesting for an algorithmic-trading platform.

Online inferenceReinforcement learningRisk controls

Inside the work

The challenge

Connect real-time order-book and alternative-data feeds to dependable, low-latency model decisions.

My contribution

Set the platform architecture, combining online inference, Redis caching, Bayesian-updated RL agents, and distributed Monte Carlo backtesting. Introduced design reviews, runbooks, and Grafana observability.

The outcome

Delivered the architecture and operational foundations to the internal team, including statistical risk controls, automated kill switches, and hot-swap rollouts.

Read the Equinox AI B.V. work story ↗

SCIENTIFIC EVIDENCE SYNTHESIS05

Synthesa AI / 2025–Present · Venture

Scientific evidence, connected

My venture work: an LLM platform for pharmaceutical and biotech evidence synthesis, with multimodal retrieval and parallel agent workflows.

Founding ML architectMultimodal RAGvLLM

Inside the work

The challenge

Coordinate retrieval and reasoning across scientific documents, tables, and images with predictable, traceable execution.

My contribution

Designed the cloud-native LLM platform using Django, Docker, GCP, distributed queues, KV caching, and vLLM. Built parallel language and vision-language agent workflows with vector search, reranking, and policy filters.

The outcome

Contributed the ML architecture behind a platform whose published validation screened 270,626 abstracts, with 100% sensitivity, 99.4% specificity, and 91.7% less manual-review workload.

Read the Synthesa AI work story ↗

02 / EXPERIENCE

From the lab.
Into the world.

Research, engineering, and technical leadership — with ownership all the way through.

Netherlands01 / 06

eBayAmsterdam

52.37° N
4.90° E

eBay

2026–Present

Senior Applied Researcher, Ads

Building multi-agent systems from scratch for recommendations, advertising-opportunity discovery, and internal workflow automation.

Amsterdam, Netherlands

Moonvalley

2024–2026

Member of Technical Staff, Foundational AI

Large-scale distributed training with model sharding, tensor and data parallelism; text-to-image/video R&D, preference optimization, and synthetic-data pipelines.

London · Remote

Synthesa AI

2025–Present · Venture

Founding ML Architect

Designed the LLM platform and multimodal RAG architecture for pharmaceutical and biotech evidence synthesis, using parallel agentic workflows.

Remote

Equinox AI B.V.

Jun–Sep 2024

Consulting ML Architect

Architected a low-latency trading platform with online inference, reinforcement-learning agents, distributed backtesting, and operational risk controls.

The Hague, Netherlands

University of Warsaw

2021–2026

Doctoral Researcher & Teaching Assistant

Research on reasoning and robustness in language and vision-language models. Taught NLP and visual recognition labs and mentored student research teams.

Warsaw, Poland

DeepSea Technologies

2018–2021

ML Engineer / Technical Lead

Led a five-engineer team working on anomaly detection, missing-data modeling, weather-aware routing, and predictive maintenance over sparse and corrupted vessel time series. Served 20M predictions daily.

Athens, Greece

03 / RESEARCH & IDEAS

But does it
actually reason?

A recurring question in my work: where does
pattern matching end and reasoning begin?
All publications on Google Scholar ↗
Read the research notes ↗

PhD dissertation · 202601

Visual and Language Reasoning in Deep Learning Models ↗

Doctoral research on visual reasoning, code generation, and constructive geometry.

EMNLP Findings · 202402

Beyond Lines and Circles: Unveiling the Geometric Reasoning Gap in Large Language Models ↗

Geometric reasoning, internal dialogue, and role-specialized multi-agent systems.

ACL Findings · 202303

A Simple, Yet Effective Approach to Finding Biases in Code Generation ↗

How function names, specifications, and examples influence generated code. Preprint: 2022.

ICLR · 202204

Measuring CLEVRness: Black-box Testing of Visual Reasoning Models ↗

Adversarial evaluation of visual reasoning without access to model internals.

IJCNN · 202105

MAIN: Multihead-Attention Imputation Networks ↗

Modeling missingness with attention to improve downstream predictions on incomplete data.

Technologies · 202006

A TensorFlow Extension Framework for Optimized Generation of Hardware CNN Inference Engines ↗

Turning TensorFlow models into configurable hardware inference engines.

IEEE MOCAST · 201907

TF2FPGA: A Framework for Projecting and Accelerating Tensorflow CNNs on FPGA Platforms ↗

Mapping and optimizing convolutional neural networks for FPGA acceleration.

04 / THE HUMAN IN THE LOOP

Hi again.
I’m Spyros.

Based in Amsterdam.
Greek roots. A curious mind.

I like the space between “what if?” and “it works.”

My work combines applied AI research, machine learning engineering, and technical leadership: multimodal generative AI, agentic systems, and production ML. I’ve built systems from scratch, led a five-engineer team, and taught master’s-level NLP and visual recognition labs.

At the University of Warsaw, my doctoral research focused on visual and language reasoning in deep learning models. That question still shapes how I build: understand what a model can do, find where it fails, and make the whole system better.

Education

PhD, Computer ScienceUniversity of Warsaw · 2026

MSc, Data ScienceAthens University of Economics and Business · 2021

BSc & MEng, Electrical and Computer EngineeringNational Technical University of Athens · 2018

Research & modeling

Multimodal generative AI · Multi-agent systems · Alignment · Reinforcement learning · RAG · Evaluation & robustness

ML systems

PyTorch · TensorFlow · Transformers · vLLM · Ray · ONNX · MLflow · Weights & Biases · Spark · Dask

Engineering & infrastructure

Python · SQL · Docker · Kubernetes · GCP · AWS · FastAPI · PostgreSQL · Redis · FAISS · Grafana

Greek Native/ English C2/ German B1

Spyros Mouselinos.

Research instincts. Real-world outcomes.

From signals to decisions

The challenge

My contribution

The outcome

Training generative models at scale

The challenge

My contribution

The outcome

Making sense of imperfect vessel data

The challenge

My contribution

The outcome

From market data to controlled execution

The challenge

My contribution

The outcome

Scientific evidence, connected

The challenge

My contribution

The outcome

From the lab. Into the world.

eBay

Moonvalley

Synthesa AI

Equinox AI B.V.

University of Warsaw

DeepSea Technologies

But does it actually reason?

Visual and Language Reasoning in Deep Learning Models ↗

Beyond Lines and Circles: Unveiling the Geometric Reasoning Gap in Large Language Models ↗

A Simple, Yet Effective Approach to Finding Biases in Code Generation ↗

Measuring CLEVRness: Black-box Testing of Visual Reasoning Models ↗

MAIN: Multihead-Attention Imputation Networks ↗

A TensorFlow Extension Framework for Optimized Generation of Hardware CNN Inference Engines ↗

TF2FPGA: A Framework for Projecting and Accelerating Tensorflow CNNs on FPGA Platforms ↗

Hi again. I’m Spyros.

Education

Research & modeling

ML systems

Engineering & infrastructure

Spyros
Mouselinos.

Research instincts.
Real-world outcomes.

From the lab.
Into the world.

But does it
actually reason?

Hi again.
I’m Spyros.