Original deep learning model (SERAPH) integrated into a full-stack protein platform that calculates structural impact diffs and generates natural-language explanations.
Building production AI systems, not just notebooks.
Track record of designing custom deep learning architectures, multi-agent pipelines, and full-stack software deployed in real-world environments.
Featured Systems
MutantScope & SERAPH
Architecture: ESM2 + Conv1D + BiLSTM trained on CullPDB benchmark dataset.
Pipeline: Next.js 15 frontend fetching structures directly from AlphaFold DB.
0.00%
Q3 Accuracy Benchmark
AI Business Analyzer
Autonomous multi-agent research pipeline generating structured business intelligence reports from a single prompt via live web search.
5-Agent Orchestration: Planner, Researcher, Aggregator, Reporter & Summarizer.
Outputs: Dynamic SWOT matrix, PDF export engine & SQLite persistent memory.
0
Autonomous Pipeline
Neural Network from Scratch
Fully modular neural network framework implemented strictly from mathematical first principles with zero external ML framework dependencies.
Engine: Hand-coded forward/backprop, Adam optimizer, BatchNorm & Dropout.
Design: Composable layer abstractions mirroring PyTorch functional API architecture.
0.00%
Test Accuracy on MNIST
Capabilities & Stack
ML & AI Core
Web & Systems
Tools & Languages
Let's build something significant.
Send a message below to trigger my immediate automated response system and establish direct contact.
Thank you! Your message has been sent successfully. I will get back to you soon.