Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I’m Sepehr Rezaee, a senior AI architect focusing on agentic LLM systems, retrieval-augmented generation, and safety-first platforms. Over 5+ years I’ve designed, evaluated, and operated production multi-agent services with SLOs, error budgets, and cost/latency controls. I’m proficient in Python, Docker/Kubernetes, and LangChain/LangGraph/LlamaIndex, and I’ve led architecture governance, standardization, and cross-functional teams across SaaS and enterprise workloads. My research grounding includes peer-reviewed publications in ICCV/NeurIPS. I’m passionate about turning research into robust production systems, building reusable platform capabilities, and driving measurable business outcomes through reliable AI services and strong safety practices.…Hi, I’m Sepehr Rezaee, a senior AI architect focusing on agentic LLM systems, retrieval-augmented generation, and safety-first platforms. Over 5+ years I’ve designed, evaluated, and operated production multi-agent services with SLOs, error budgets, and cost/latency controls. I’m proficient in Python, Docker/Kubernetes, and LangChain/LangGraph/LlamaIndex, and I’ve led architecture governance, standardization, and cross-functional teams across SaaS and enterprise workloads. My research grounding includes peer-reviewed publications in ICCV/NeurIPS. I’m passionate about turning research into robust production systems, building reusable platform capabilities, and driving measurable business outcomes through reliable AI services and strong safety practices.

Sepehr Rezaee

Architect, AI Engineer, Data Scientist, +6





Hi, I’m Sepehr Rezaee, a senior AI architect focusing on agentic LLM systems, retrieval-augmented generation, and safety-first platforms. Over 5+ years I’ve designed, evaluated, and operated production multi-agent services with SLOs, error budgets, and cost/latency controls. I’m proficient in Python, Docker/Kubernetes, and LangChain/LangGraph/LlamaIndex, and I’ve led architecture governance, standardization, and cross-functional teams across SaaS and enterprise workloads. My research grounding includes peer-reviewed publications in ICCV/NeurIPS. I’m passionate about turning research into robust production systems, building reusable platform capabilities, and driving measurable business outcomes through reliable AI services and strong safety practices.…Hi, I’m Sepehr Rezaee, a senior AI architect focusing on agentic LLM systems, retrieval-augmented generation, and safety-first platforms. Over 5+ years I’ve designed, evaluated, and operated production multi-agent services with SLOs, error budgets, and cost/latency controls. I’m proficient in Python, Docker/Kubernetes, and LangChain/LangGraph/LlamaIndex, and I’ve led architecture governance, standardization, and cross-functional teams across SaaS and enterprise workloads. My research grounding includes peer-reviewed publications in ICCV/NeurIPS. I’m passionate about turning research into robust production systems, building reusable platform capabilities, and driving measurable business outcomes through reliable AI services and strong safety practices.

Available to hire

Hi, I’m Sepehr Rezaee, a senior AI architect focusing on agentic LLM systems, retrieval-augmented generation, and safety-first platforms. Over 5+ years I’ve designed, evaluated, and operated production multi-agent services with SLOs, error budgets, and cost/latency controls. I’m proficient in Python, Docker/Kubernetes, and LangChain/LangGraph/LlamaIndex, and I’ve led architecture governance, standardization, and cross-functional teams across SaaS and enterprise workloads. My research grounding includes peer-reviewed publications in ICCV/NeurIPS.

I’m passionate about turning research into robust production systems, building reusable platform capabilities, and driving measurable business outcomes through reliable AI services and strong safety practices.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Language

Persian

Fluent

English

Advanced

Work Experience

Senior LLM Engineer at AIR Property

August 1, 2025 - Present

Architected and shipped production LLM services (RAG + agents), owning model selection, agent/prompt design, eval harnesses, and fallback trees; improved answer quality within strict latency and cost budgets. Introduced an architecture blueprint for agent services and a governed toolchain (registries, policies, approvals), increasing reliability and reusability. Built modular APIs with tests and dashboards for SLOs, error budgets, safety metrics; collaborated with product/security to align thresholds and incident response. Led data curation, vector caching, and inference optimization to stabilize throughput under peak load; managed Kubernetes capacity planning and rollouts.

AI Engineer at PropTy Global

August 1, 2024 - September 1, 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; standardized routing/hand-off patterns. Productionized on Docker/Kubernetes with sub-100ms API paths for critical endpoints; instrumented with Prometheus/Grafana and centralized logging. Closed the loop to live KPIs with automated evaluation/feedback for continuous improvement and drift monitoring.

Chief AI Officer & Multi-Agent Architect at Novel Mind Scientist

October 1, 2022 - September 1, 2025

Led delivery of LLM-powered agents across SaaS/health/education, integrating text/vision and knowledge graphs with measurable SLAs. Scaled multi-agent orchestration (LangChain, Celery) and automated business processes; established design reviews, eval protocols, and knowledge transfer. Drove engineering standards (docs, ADRs, onboarding guides) to accelerate adoption and reduce operational risk.

Research Intern at Safe & Generative AI Mathis Lab, EPFL

May 1, 2025 - September 30, 2025

Co-authored ICCV 2025 (accepted) paper on DISTIL: data-free diffusion-based trigger inversion for Trojaned models; 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 for safer AI. Developed zero-shot, data-free defenses and ran large-scale benchmarks; published best practices for robust evaluation.

Research Assistant — Agentic AI & Security at Sharif University & Shahid Beheshti University

January 1, 2023 - January 1, 2025

Prototyped secure agentic ML pipelines: RAG, routing/hand-off, memory management. Published/submitted work to NeurIPS/ICCV on agent security, evaluation, and optimization; mentored junior engineers.

Project Manager, Agentic ML SaaS at NovaVira

March 1, 2023 - February 1, 2024

Delivered Django-based agentic recommender (LangChain, GCP, Docker) with hybrid search and automated workflows. Ran Agile CI/CD to accelerate iteration on agent architectures and platform reliability.

AI Engineer at Agentic Systems PropTy Global

August 1, 2024 - September 1, 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; standardized routing/hand-off patterns. Productionized on Docker/Kubernetes with sub-100ms API path for critical endpoints; instrumented with Prometheus/Grafana and centralized logging. Closed the loop to live KPIs with automated evaluation/feedback for continuous improvement and drift monitoring.

Research Intern at Safe & Generative AI Mathis Lab, EPFL (Switzerland)

May 1, 2025 - September 1, 2025

Co-authored ICCV 2025 DISTIL: data-free diffusion-based trigger inversion for Trojaned models; 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 for safer AI. Developed zero-shot, data-free defenses and ran large-scale benchmarks; published best practices for robust evaluation.

Research Assistant — Agentic AI & Security at Sharif Univ. & Shahid Beheshti Univ.

January 1, 2023 - January 1, 2025

Prototyped secure agentic ML pipelines: RAG, routing/hand-off, memory management; published/submitted work to NeurIPS/ICCV on agent security, evaluation, and optimization.

Project Manager at NovaVira

March 1, 2023 - February 1, 2024

Delivered Django-based agentic recommender (LangChain, GCP, Docker) with hybrid search and automated workflows; ran Agile CI/CD to accelerate iteration on agent architectures and platform reliability.

AI Engineer at PropTy Global (Agentic Systems)

August 1, 2024 - September 1, 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; standardized routing/hand-off patterns. Productionized on Docker/Kubernetes with sub-100ms API path for critical endpoints; instrumented with Prometheus/Grafana and centralized logging. Closed the loop to live KPIs with automated evaluation/feedback for continuous improvement and drift monitoring.

Research Intern at Mathis Lab, EPFL

May 1, 2025 - September 1, 2025