class TusharKhete:
def __init__(self):
self.role = "Generative AI Engineer"
self.location = "🌍 Global · Remote"
self.focus = ["LLMs", "Multi-Agent Systems", "RAG", "Fine-tuning"]
self.exploring = ["Agentic AI", "Model Alignment", "Reasoning Models"]
self.philosophy = "Involve me and I learn."
def current_stack(self):
return "LangGraph • CrewAI • PyTorch • HuggingFace • vLLM"
def say_hi(self):
print("Let's build something intelligent together 🚀")
|
Security-first multi-agent swarm framework with tri-layer self-improving memory and pluggable skill architecture.
|
Root LLM gets a Python sandbox and delegates focused sub-questions to cheaper sub-models, then synthesizes — with a fintech 10-K demo.
|
|
Fine-tuned DeepSeek R1 on a custom mental-health dataset for empathetic reasoning.
|
Multi-agent system analyzing investment risk with collaborating specialist agents.
|
Because good engineering, like good chess, is all about thinking several moves ahead.
⭐️ From tushark01 — turning research into production AI.