class Sobhan:
def __init__(self):
self.role = "CS Undergrad @ JIIT Noida (2023–2027)"
self.focus = ["Agentic AI", "RAG Systems", "Backend Engineering"]
self.currently_building = "Cloud observability platform @ LTM "
self.fun_fact = "Coding while listning to soft music with a cup of coffee - that's my idea of a perfect evening."
def say_hi(self):
return "Let's build something intelligent together!"| Software Engineer Intern @ LTM (Jun 2026 – Present) | Building a cloud-ready observability platform — Prometheus, Grafana & OpenTelemetry dashboards keeping infra reliable as it ships to the cloud. | |
| **Software Engineer-AI/ML Intern @ Trafiksol IT Technologies (Apr 2025–May 2026) | Built the LangGraph agent behind an AI traffic-signal system. It recommends timing changes using forecasting, policy-RAG and anomaly services, with a non-LLM safety check on every action and human escalation when it lacks grounding. Also fine-tuned and quantized a small language model for an internal company chatbot. |
|
ReWOO · MCP · LSTM · JWT Modular agent system (Planner → Worker → Solver → Refiner) with dynamic MCP tool orchestration. ⚡ Sub-3s responses ✅ 80% task completion rate |
FAISS · BM25 · KG · FastAPI Combined vector search, BM25 & knowledge graphs across 4 medical policy docs. 🎯 87.6% groundedness score 🤖 LLM-as-a-judge validation |
CV · Neural Nets · Flask MultiSSL + U-KAN segmentation on 3.4GB+ of Sentinel-2 satellite imagery. 📍 15+ restoration sites found 🛰️ Geospatial ML pipeline |
Languages
AI / ML & LLMOps
Backend & Systems
Data & Storage
Cloud & Observability
🔧 Backend system design for AI-driven apps · 📦 Scalable ML deployment · 🧠 Production-ready agent architectures
