Skip to content
View SepehrRezaee's full-sized avatar
🌍
Hello World!!
🌍
Hello World!!
29 followers · 72 following

Achievements

Achievement: QuickdrawAchievement: Pull Shark

Achievements

Achievement: QuickdrawAchievement: Pull Shark

Highlights

  • Pro

Block or report SepehrRezaee

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
SepehrRezaee/README.md

Hi there 👋

I'm Sepehr Rezaee, a Senior AI Architect specializing in agentic LLM systems, retrieval-augmented generation, and safety-first platforms with 5+ years of experience designing and operating production multi-agent services.

My focus areas:

  • AI Architecture & Multi-Agent Systems: Designing scalable production multi-agent services with SLOs, error budgets, and cost/latency controls
  • RAG Platforms & Safety: Building retrieval-augmented generation systems with enterprise-grade governance and security
  • MLOps at Scale: Orchestrating reliable AI systems with Docker/Kubernetes, monitoring, and observability
  • AI Security & Research: Translating research on model security and evaluation into production practices

See more in my CV.

🚀 Experience

  • AI Engineer, Agentic Systems @ PropTy Global (Dubai, 2024–2025)

    • Architected multi-agent systems (LangChain + custom RAG) for autonomous recommendations/decisions
    • Achieved 85%+ end-to-end task completion in business workflows
    • Implemented agent-to-agent protocols and memory for context-aware planning and goal execution
    • Productionized on Docker/Kubernetes with sub-100ms API path for critical endpoints

  • Chief AI Officer & Multi-Agent Architect @ Novel Mind Scientist (Tehran, 2022–2025)

    • Led delivery of LLM-powered agents across SaaS/health/education with measurable SLAs
    • Scaled multi-agent orchestration (LangChain, Celery) and automated business processes
    • Drove engineering standards (docs, ADRs, onboarding guides) to accelerate adoption

  • Research Intern @ Mathis Lab, EPFL (Switzerland, 2025)

    • Co-authored ICCV 2025 accepted paper on DISTIL: data-free diffusion-based trigger inversion for Trojaned models
    • Achieved new SOTA on BackdoorBench (+7.1% acc) and object-detection scanning (+9.4%)
    • Built latent-diffusion pipelines with classifier-guided feedback to expose adversarial vulnerabilities

🛠️ Core Skills & Technologies

  • Agent Orchestration: LangChain, LangGraph, LlamaIndex, Multi-Agent Systems, RAG, SPAR (Sense-Plan-Act-Reflect), Prompt Engineering
  • Programming: Python (expert), C++, Java, C#
  • Infrastructure: Docker (expert), Kubernetes (expert), AWS (SageMaker, EC2), GCP, Azure
  • Databases/Vector: Pinecone, Weaviate, Chroma, PostgreSQL (pgvector), MongoDB, Redis
  • LLMs: OpenAI GPT-3/4, Anthropic Claude, Google Gemini, HF Transformers
  • MLOps & Services: MLflow, Airflow, Celery, Prometheus, Grafana, ELK, FastAPI, Flask, REST/GraphQL
  • Other: Knowledge Graphs (Neo4j), Multimodal AI, Speech/Text Interfaces, SaaS Architecture

📚 Publications

  • DISTIL: Data-Free Inversion of Suspicious Trojan Inputs via Latent Diffusion (ICCV 2025, accepted)
  • Scanning Trojaned Models Using Out-of-Distribution Samples (NeurIPS 2024, accepted)
  • Comparison of Pre-Training and Classification Models for Early Detection of Alzheimer's Disease Using MRI (I4C 2023)

🏆 Awards & Honors

  • Best Ideator Award: National Young Scientists Festival (2023)
  • Top 0.2% National Entrance Exam: Placed 352nd out of ~150,000 (2020)

💬 Let's Connect!

Profile views SepehrRezaee's GitHub stats Top Langs

Pinned Loading

  1. omani_therapist omani_therapist Public

    Python 1

  2. An-implementation-of-Semantic-IMDB-Regression-using-TensorFlow An-implementation-of-Semantic-IMDB-Regression-using-TensorFlow Public

    Jupyter Notebook 3

  3. ImageProcessing_CourseOf_ShahidBeheshtiUniversity ImageProcessing_CourseOf_ShahidBeheshtiUniversity Public

    This repository is about the assignments of Image Processing course of Shahid Beheshti University with its implementations

    Jupyter Notebook 2 1

  4. deep-learning-symbolic-mathematics deep-learning-symbolic-mathematics Public

    Forked from elia-mercatanti/deep-learning-symbolic-mathematics

    Discussion and test of the first successful approach to solving symbolic mathematics problems through the use of neural networks, proposed for the first time by two Facebook researchers, Guillaume …

    Jupyter Notebook 2

  5. neural-netwroks-from-scratch neural-netwroks-from-scratch Public

    Forked from alirezaafzalaghaei/neural-networks-from-scratch

    implementing some neural networks in python

    Jupyter Notebook 2

  6. NN-Fall-2022 NN-Fall-2022 Public

    This repository is about the neural netowrk course that helding in Shahid Beheshti University

    Jupyter Notebook 1