I build systems that turn messy input into something a machine can execute — LLM pipelines, agents, and the infrastructure around them. Springer-published in deep learning; before that I wrote real-time control code for aero-engines at DRDO.
Most of what's below was built from scratch on purpose. A database engine without an ORM, a neural network without a framework, a CRDT editor without a backend that can read your data. It's the fastest way I know to actually understand a thing.
AI Intern · Infineon Technologies — Bengaluru, India · Mar 2026 – Present
Building an LLM system that converts multimodal input (text, sketches, diagrams) into executable BPMN 2.0 workflow models, deployed on a Red Hat OpenShift–based enterprise hybrid cloud.
| QueryForge · live ↗ A SQL database engine written from scratch — B+tree storage, write-ahead logging, and a Volcano-model executor. Python TypeScript B+tree WAL |
Synthesis · live ↗ An autonomous multi-agent research engine that plans, searches, and synthesises a cited answer. TypeScript LLM agents RAG |
| Tracewave Real-time streaming anomaly detection — SSE ingestion, one-second tumbling windows, and an ensemble of three online detectors with an explainable "why". Python Redis Streams TimescaleDB Prometheus |
Synapse · live ↗ A local-first, end-to-end-encrypted collaborative canvas. The relay only ever sees ciphertext. Next.js Yjs CRDT AES-256-GCM |
| NeuroForge · live ↗ A neural network — forward pass, backprop, optimizers — hand-written in vanilla JS with zero dependencies. Gradient check agrees to ~2e-10. JavaScript Backprop Canvas |
MeetEase · live ↗ AI-native video conferencing with scheduling, recordings, and real-time rooms. Next.js TypeScript Stream SDK |
More on the portfolio ↗ · everything else lives in repositories ↗
Lung disease detection from chest X-rays — a hybrid ResNet50 architecture reaching 97.18% accuracy. Published by Springer.
| Languages | Python · TypeScript · JavaScript · SQL · C++ |
| AI / ML | PyTorch · TensorFlow · LangChain · scikit-learn · Pandas |
| Backend & data | FastAPI · Node.js · PostgreSQL · Redis · MongoDB · TimescaleDB |
| Frontend | React · Next.js · Tailwind |
| Platform | Docker · Kubernetes · OpenShift · AWS · GitHub Actions · Grafana |
- B.Tech, Computer Science (Data Science) · Christ University, Bengaluru · CGPA 8.9/10
- 1st place, ICE Tech 2025 — national hackathon
- Department Head, CSE Student Body — led initiatives for 400+ students · Student Council Head
- Research Intern, DRDO — algorithms and test automation for aero-engine control systems
Open to conversations about LLM systems, agents, and developer tooling.
charanreddy.dev · LinkedIn · charanreddychanda@gmail.com