MS Computer Engineering @ Virginia Tech | Software Engineer | ML Researcher

I build AI infrastructure, train LLMs to reason better, and ship systems that scale. Working on open-source projects with 300+ active users.

class DarshanNere:
    def __init__(self):
        self.role     = "MS Computer Engineering @ Virginia Tech"
        self.graduating = "May 2026"
        self.interests = [
            "AI Agent Infrastructure & Observability",
            "Reinforcement L
            "LLM Security & Multi-Agent Systems",
            "Full-Stack & Sy
            "Human-Computer Interaction",
        ]
        self.current_work = {
            "building":  "Tracia platform",
            "open_src":  "ObservAgent — 300+ active users",
            "research":  "LLfeedback (FlexGuard)",
            "teaching":  "Graduate TA @ VT (ML + C++, 180 students)",
        }

    def say_hi(self):
        print("Open to full-time roles in SWE / ML Eng / AI Infra — May 2026 🚀")

me = DarshanNere()
me.say_hi()

• 🤖 Open-Sourcing ObservAgent: Making Claude Code Sessions Actually Observable — Mar 2026
• 🐳 Running OpenClaw Safely with Docker — Feb 2026
• 🏋️ FlexGuard: Designing On-Body Feedback to Prevent Injuries in Strength Training — Jan 2026
• 🔒 A Cross-Layer Survey of Security for Large Language Models — Dec 2025

➡️ Read more on darshannere.com/blog

ai_infrastructure:
  - Tracea: team AI agent observability — FastAPI, React, YAML rules, MCP server
  - ObservAgent: open sourcee integration
  - OpenClaw Docker: safe containerization for autonomous AI agents

ml_research:
  - GRPO on Qwen2.5-3B: 71% seline (-33% VRAM)
  - Medical CoT + GRPO: +11% accuracy over SFT on 90k+ samples (single T4)
  - LLM Security Survey: Neus security layer
  - FlexGuard: arXiv:2509.18662, HCI wearable feedback design space

systems:
  - 10M queries/second searc space indexing
  - Municipal software serving 50,000+ citizens (Pune Municipal Corporation)
  - 40% reduction in server on + connection pooling

const currentlyWorkingOn = {
  "🤖 building": [
    "Tracea — AI agent obserrules + AI-powered RCA",
    "ObservAgent — open source monitoring for AI agent sessions",
  ],
  "🔬 research": [
    "GRPO / DPO for LLM reas
    "Multi-agent system security and attack surface analysis",
    "Wearable feedback systexGuard)",
  ],
  "🎓 learning": [
    "Advanced RL techniques (GRPO, DPO, PPO)",
    "MCP server design and a
    "Production-scale LLM inference optimization",
  ],
  "💼 seeking": [
    "Software Engineering —
    "ML Engineering / AI Infrastructure roles",
    "Research collaborations",
  ],
};

I love to build and break things — from AI observability platforms to fine-tuning LLMs on consumer hardware. I hard, let's talk.

Open to:

• 💼 Full-time SWE / ML Eng / AI Infra roles (available May 2026)
• 🤝 Research collaborationscurity
• 💡 Open source contributions and interesting technical discussions

Reach me:

• 📧 darshan.nere2@gmail.com
• 💼 linkedin.com/in/darshan-nere
• 🌐 darshannere.com
• 🐦 @DarshanNere

• 🤖 Built an observability platform used by 300+ AI developers worldwide
• 🏗️ Shipped software servinnicipal Corporation
• 🧠 Fine-tuned Qwen2.5-3B to 71% on GSM8K — on a single consumer GPU
• ☕ Fueled by coffee, curiosity, and unreasonably hard problems

✨ Thanks for visiting! Let's build something that matters.