Maede Jalali
Senior AI Engineer
Designing modular agentic architectures, multi-agent orchestration layers, and conversational AI platforms with evaluation-driven quality and safety guardrails.
01
About
Senior AI Engineer with 6+ years building production Python systems and 3+ years in applied GenAI. I design and ship end-to-end agentic applications: stateful multi-agent orchestration (LangGraph), production RAG pipelines, and evaluation frameworks with safety guardrails.
I deploy on Azure and Docker/Kubernetes, and work closely with product and engineering teams to take problems from architecture through to measurable impact. I thrive in ownership-heavy environments where technical decisions translate into real outcomes.
02
Experience
Building production Python AI agents orchestrating data, tools, and multi-step decisions across CRM and outreach systems.
- —Built production Python AI agents orchestrating data, tools, and multi-step decisions across CRM and outreach systems using LangGraph/LangChain, structured outputs, and RAG.
- —Rebuilt workflows into a modular orchestration layer, cutting latency 15% and costs 45%.
- —Automated lead enrichment (LinkedIn, company metadata, address, headcount), saving ~125 hours / 1,000 leads (~$5k per batch).
- —Built an evaluation and experimentation loop across models (OpenAI, Google) and prompts to select optimal configurations, integrated into CI/CD.
- —Added guardrails (Pydantic validation, hallucination checks, injection defenses, PII redaction) and participated in Google Security Assessment for data privacy compliance.
- —Deployed and operated AI applications using Docker and Kubernetes; integrated agents with enterprise systems, prioritising reliability and debuggability through tracing and analytics.
- —Mentored backend engineers on agentic patterns, code quality, and best practices for production AI systems.
- —Partnered with sales and product to ship automation improving outbound results, including a +16% open rate.
04
Skills
Agentic Frameworks
- LangGraph
- LangChain
- MCP
- Agentic Design Patterns
- Multi-Agent Communication
- Autonomous Planning & Reasoning
LLM & RAG
- OpenAI API
- Anthropic API
- HuggingFace
- RAG Pipelines
- Embeddings
- Vector Search
- ChromaDB
- Prompt Engineering
Evaluation & Safety
- LLM Judge
- OpenAI Evals
- BLEU / ROUGE / BERTScore
- Guardrails & Pydantic
- Hallucination Defenses
- PII Redaction
- Observability & Tracing
Backend & Cloud
- Python
- Django
- FastAPI
- Celery
- Async
- SSE Streaming
- Azure AI
- GCP
- Docker
- Kubernetes
- CI/CD
Data & Infrastructure
- PostgreSQL
- Redis
- Vector Databases
- Metabase
- Amplitude
- Pytest
Governance & Compliance
- AI Safety Frameworks
- Data Privacy
- Injection Defenses
- Audit Logging
03
Projects
AI Travel Assistant
End-to-end AI-powered Telegram bot enabling users to book flights and hotels via natural language. Reduced booking time from 10 minutes to under 1 minute.
IoT Airport Lighting System
Mobile-based control system for airport apron lighting. Operators adjust settings in 10 seconds via an intuitive real-time interface.
05
AI Assistant
Ask my AI assistant anything about my experience, skills, or background.
06
Contact
Interested in working together? Let's connect.
majj2000.mail@gmail.com
Phone
+98 (933) 426-5571
Location
Tehran, Iran
linkedin.com/in/maede-jj