Open to opportunities
Senior Software Engineer · Knowledge Graphs, Graph ML & AI Infrastructure
Software engineer with 6+ years building knowledge graphs and distributed systems. Former Neo4j SME on credit-risk graph platforms at Credit Suisse and AI Engineer at PhysicianX, where I worked on GNN-based recommendations and production RAG/multi-agent systems. MS in Computer Science from the University at Buffalo, with two 2026 publications (ASME IDETC-CIE, AIAA AVIATION) on graph learning for combinatorial optimization.
New York, NY
An in-memory, GQL-compliant graph database written in pure Java. Built for ephemeral analytics and data pipelines, zero-dependency embedded processing, and fast integration testing in CI/CD where spinning up Docker slows the pipeline down.
A real-time applied-ML recommendation engine using Neo4j graph mining and Python to run risk and sentiment assessments on highly vulnerable assets.
A deep learning framework that uses graph learning to optimize active tether-net systems for space debris capture, with scalable data structures and custom graph mining algorithms for high-dimensional combinatorial optimization spaces. Published at AIAA AVIATION 2026.
Designed, deployed and evaluated a real-time autonomous trading system that executed high-frequency order strategies during a six-month live experiment on NASDAQ using paper currency.
Fine-tuned a SymphonyNet Transformer with a permutation-invariant language model that learns music as symbols, generating long-form symphonies from short input snippets to help reduce anxiety.
Jun 2025 – May 2026
PhysicianX | Buffalo, NY
• Designed, ran and evaluated experiments training Graph Neural Networks (GNNs) on Neo4j social graphs, applying graph mining to improve user-facing recommendations. • Built a production RAG architecture and multi-agent orchestration system with Mastra Workflows and GraphQL, with strict context assembly and anti-hallucination guardrails. • Led UI design and built the full-stack app for a proprietary chat-based tool and an SEO-compliant data aggregator (React, Node.js). • Managed end-to-end ML infrastructure on GCP: continuous model deployment, evaluation and data processing for high-throughput microservices. • Implemented CI/CD frameworks that cut deployment latency by 75%.
Jan 2025 – May 2026
University at Buffalo | Buffalo, NY
• Research on graph learning for mixed combinatorial-nonlinear optimization. • Developed a deep learning framework using graph learning to optimize active tether-net systems for space debris capture, with custom graph mining algorithms for high-dimensional combinatorial search spaces. • Work led to two accepted 2026 papers: ASME IDETC-CIE 2026 (learning-based directed graph abstraction for order-preserving search) and AIAA AVIATION 2026 (active tether-net design with graph-learning-aided optimization).
Sep 2019 – Jul 2024
Credit Suisse AG | Pune, India
• Designed applied ML for a fraud detection engine using graph mining, clustering and community detection over large-scale knowledge graphs. • Subject Matter Expert for graph modeling; architected distributed data pipelines handling continuous data flow from 20 global locations. • Optimized ETL processes for an 80% increase in data processing efficiency. • Led Java (8 → 11 → 17) upgrades and Neo4j Enterprise (3 → 4 → 5) data migrations with zero downtime. • Built a community detection system for the knowledge graph and event-driven microservices for responsiveness and scale. • Drove design and code reviews with cross-functional stakeholders for accuracy, testability and engineering best practice.
University at Buffalo | Buffalo, NY | Aug 2024 – May 2026
Neo4j
Find Gishnu Madhu elsewhere on the web.