Software Developer
Arnav Panigrahi
Backend systems, fullstack applications, and AI/ML pipelines -- shipped to production.
Background
About
I build backend services, fullstack applications, and AI/ML pipelines with a focus on shipping reliable software to production. My work spans FastAPI microservices on Google Cloud, LangChain-based AI agents for enterprise workflows, and multi-modal ML systems combining computer vision, NLP, and voice synthesis.
At AllCheer, I've delivered measurable impact -- a 60% reduction in document processing time, $20K in annual cost savings, and 92% accuracy on an NER pipeline for clinical data extraction. I thrive at the intersection of backend engineering and applied AI, where system design meets real-world constraints.
I hold an M.S. in Computer Science from UC Riverside, where I focused on distributed systems and machine learning. I'm currently seeking software engineering roles in backend systems and applied AI.
Experience
Where I've worked.
Software Developer
AllCheer- Built a LangChain AI agent automating ROI processing -- authorization checks, compliance validation, and document generation -- cutting processing time by 60% and saving $20K/year.
- Engineered a containerized FastAPI service emulating authenticated RethinkBH sessions for real-time clinical record sync, bypassing 24-48hr public API latency.
- Developed a React + FastAPI NER pipeline with OpenAI for structured data extraction from therapist notes at 92% accuracy.
Software Developer Intern
AllCheer- Designed a Python logistics engine on Google Cloud with Google Maps APIs for multi-stop route optimization and large-scale staff scheduling.
- Led backend R&D and API schema design for ROI automation, prototyping in Make.com and translating to production service architecture.
Capabilities
Tools I use most.
Languages & Frameworks
- Python
- TypeScript
- JavaScript
- C#
- .NET
- FastAPI
- React
- Astro
AI & Data Systems
- LangChain
- LangGraph
- PyTorch
- Vector Search
- Embeddings
- Prompt Engineering
Infrastructure & Delivery
- Google Cloud
- Docker
- AWS
- REST APIs
- PostgreSQL
- SQL
- Git
Selected Work
Projects
bedtime.ai
Extended a multi-modal AI storytelling system that turns children's drawings into narrated stories. Four models orchestrated in a single pipeline: EfficientNet-B0 for sketch classification (~70% accuracy on children's drawings), OpenAI GPT for story generation, Coqui XTTS v2 for voice cloning, and a touch-enabled React drawing canvas. Includes 35 backend tests, 19 frontend tests, and Docker Compose deployment.
- 4 models orchestrated in a single pipeline
- ~70% classification accuracy on children's drawings
- 35 backend + 19 frontend tests
Abliteration for LFM2.5
An interpretability project for LiquidAI/LFM2.5-1.2B-Instruct focused on refusal directions, weight orthogonalization, and inference-time intervention using manual PyTorch hooks. Reduced model refusal rate from 50% to ~37% without retraining.
Nier Archive
A real-time collaborative site with live cursor sharing and viewer presence, built with SvelteKit 5, Cloudflare Workers, Durable Objects for stateful WebSocket connections, KV-backed analytics, and a custom Vite markdown pipeline for blog content.
Rethink BH Automation
A containerized FastAPI service deployed on Google Cloud Run that emulates authenticated user sessions against RethinkBH endpoints to sync appointment and authorization data in real time, bypassing 24-48 hour API latency for downstream applications.
LangChain RAG
A production-ready RAG pipeline querying Star Wars scripts using LangChain, OpenAI embeddings, and Qdrant. Rebuilt from a tutorial with configurable data sources, prompt hardening against injection, source attribution, ~30x cost reduction (GPT-4o-mini), and a 28-test suite (13 mocked unit + 15 integration).
View on GitHubEnterprise ROI System
An LLM agent that automates HIPAA document release workflows -- orchestrating authorization checks, compliance validation, and document generation to reduce manual processing time by 60% and deliver $20K in annual savings.