Applied AI engineer focused on agentic systems, multimodal models, and signal intelligence.
MEng Applied Artificial Intelligence (2026), Data Engineering.
Building systems that perceive, interpret, and act.
- SentinelDNS — An applied ML system for domain risk scoring and anomaly detection, focused on extracting signal from noisy network and behavioral data under real-world constraints.
-
RegimeAlpha — LSTM + attention time-series system for identifying BTC market regimes, with drift detection and a walk-forward evaluation engine.
-
SpinOut Radar — RAG-powered deep-tech scouting system using embeddings, Supabase, and AI summaries to surface emerging spinouts and trends.
-
DeepFile — Intelligence document extraction and retrieval system combining OCR, embeddings, and retrieval-augmented QA.
-
Landflip — Geospatial pipeline for Brazilian coastal land intelligence built on OpenStreetMap data and Python.
-
ShopperOS — Cross-web personal shopping AI using taste graphs and fit-profile embeddings to improve purchase decisions.
-
LabelFlow — GenAI-native labeling platform with RAG evaluation, hallucination testing, and instant payouts.
-
Chart Vision — Deep learning study on candlestick chart prediction comparing CNN architectures; a simple 4-layer CNN achieved 0.892 AUC-ROC, outperforming ResNet, EfficientNet, and Vision Transformers.
-
BioLink — Drug–disease modeling research using BioBERT / PubMedBERT embeddings and knowledge-graph propagation.
-
Federated IDs — Research on federated learning for intrusion detection, evaluating privacy-preserving distributed training across network security datasets.
- Chasing weak signal in noisy data and turning it into leverage
- Bringing recent ML research into existing systems where it creates real value
- Iterating quickly with AI-assisted workflows to test and refine ideas
- Choosing projects where small improvements compound into real impact
- Building security-minded, end-to-end systems meant to hold up in practice
- Founder & Applied AI Engineer at Parallax Lab, applying state-of-the-art machine learning methods to extract signal from complex data and build end-to-end AI systems that improve operational and financial outcomes.
- Focus: signal systems, agent workflows, multimodal perception, decision-architecture
- Professional experience in cyber security, infrastructure, and machine learning
See my detailed professional profile on LinkedIn:
🔗 https://www.linkedin.com/in/dustinhaggett/
📫 Email: dustin@parallaxlab.ai
🌐 Portfolio: https://dustinhaggett.com
🏢 Parallax Lab: https://parallaxlab.ai
💼 LinkedIn: https://www.linkedin.com/in/dustinhaggett/