Tridib Dutta

About Tridib Dutta

Principal Data Scientist with 10+ years of experience building and deploying ML systems at enterprise scale, with deep expertise in Generative AI, Graph ML, and applied NLP. My work spans the full stack — from research and model development to production deployment and evaluation frameworks.At Absolute Security, I architected the GenAI core of a real-time AI Threat Insights platform monitoring devices, engineered Graph Neural Networks to detect relational anomalies across complex network topologies, and built production-ready RAG systems evaluated via Ragas for factual accuracy and instruction-following integrity. On the LLM side, I\'ve fine-tuned locally-hosted transformer models for domain-specific tasks and built text-to-SQL pipelines achieving ~75% accuracy on moderately complex queries.Earlier in my career I delivered NLP and ML solutions for Fortune 50 clients including Microsoft, Boeing, and Apple through Infosys, and published applied ML research out of Cornell University in agricultural economics.I hold a PhD in Mathematics and remain actively engaged in the broader ML community — including an open-source contribution to PyTorch Geometric (PyG), in collaboration with Matthew Fay, and recent applied research through the NIH All of Us Program and the DREAM Target 2035 Drug Discovery Challenge.Open to senior and principal-level AI/ML roles where deep technical ownership meets real-world impact. d••••••••@gmail.com | Seattle, WA

Experience

1. Principal Data Scientist

   Netmotion Software By Absolute

   Oct 2022 — Present · Seattle, WA, US

   AI Threat Insights PlatformConceptualized and built the GenAI core of a real-time zero-day detection pipeline, integrating predictive models to categorize network threats (e.g, port scanning) across monitored devices. Designed the foundational ML architecture prior to handoff to engineering for enterprise-scale deployment.* LLM & RAG SystemsFine-tuned locally-hosted transformer models (Huggingface) for domain-specific tasks to prevent sensitive data leakage. Built a text-to-SQL pipeline using neural intent parsing and entity linking, achieving ~75% factual accuracy on moderately complex queries. Developed a production-ready RAG system using Chroma for vector storage and semantic search; established evaluation frameworks via Ragas to measure factual accuracy and instruction-following integrity.* Graph Neural Networks for Network SecurityEngineered GNNs to model complex network topologies and detect relational anomalies undetectable by traditional ML methods. Applied findings to refine cohort analysis and strengthen enterprise security posture.* Cross-Functional LeadershipLed joint initiatives with engineering and product teams; translated complex ML concepts for non-technical stakeholders and mentored junior data scientists.

Skills

• Statistical Modeling
• Research
• C++
• C
• Programming
• Predictive Modeling
• Network Biology
• Perl
• Physics
• Data Science
• Data Visualization
• Computer Science
• Machine Learning
• R
• Algorithms
• Mac
• Sql
• Graphs
• Python
• Latex
• Parallel Computing
• Mathematics
• Mpi
• Matlab
• Data Mining
• Weka
• Fortran
• Statistics
• Science
• Data Analysis
• Unix
• Xp
• Problem Solving
• Numerical Analysis
• Bioinformatics
• Awk
• Applied Mathematics
• Mathematical Modeling
• Mathematica
• Computational Biology