> Research Assistant @ UNC Chapel Hill | MS Computer Science

> Currently working on formal verification for RISC-V architectures

---

$ cat tech_stack.txt

// Languages

// AI & ML

// Backend & Cloud

// Cryptography & Security

// Systems & Tools

---

---

$ ls ~/projects

Project	Description
Garuda	RISC-V INT8 accelerator with 7.5-9× latency reduction for attention microkernels
PSF Design Lab	AI predictability assessment toolkit with hybrid measurement (entropy + KL divergence + LLM)
TraceLens	Visual debugger for LangGraph workflows with time-travel debugging and real-time visualization
VAC	Capability-based security for AI agents with task-scoped credentials and instant revocation
Axiom	Hybrid verification framework (KLEE + TLA+) for distributed systems with 100% coverage for Raft
Serverless RAG Chatbot	Scale-to-zero RAG architecture with $0/month when idle and 30-60% cost reduction via caching

---

$ git log --author=param --oneline

Contribution	Description
pydantic-ai	Fixed race condition in parallel tool execution with targeted locking
vLLM	INT8 quantization fix for Blackwell GPUs (SM100+) with early validation and FP8 guidance
CVA6 RISC-V	Fixed interrupt encoding in RV64 + inferred latch fix - improves hardware reliability
MCP Python SDK	Resource leak fix in SSE handlers with comprehensive test suite
Anthropic Skills	Performance optimization: 90% reduction in bounding box comparisons for multi-page forms

---

$ curl blog.rss

• History of World Models: From Mental Simulators to AI’s Next Frontier
• The Molting of the Claw: Why Autonomous Agents Need a Verifiable Shell
• The Paradox of Choice: A Human’s Guide to Selecting the Right AI Model

---

$ tail -f activity.log

1. 🗣 Commented on #3729 in pydantic/pydantic-ai
2. 🗣 Commented on #3729 in pydantic/pydantic-ai
3. 🎉 Merged PR #3259 in emmabostian/developer-portfolios

---

Made while doing lot of snacking ;)

param@github: ~ $ exit