Senior .NET / Backend / Enterprise / AI Engineer

I design and build secure, reliable backend systems with C#, ASP.NET Core, SQL Server, PostgreSQL, Redis, Docker, and OpenTelemetry, with a focus on enterprise workflows, multi-tenant architecture, data integrity, observability, and AI-enabled applications.

My engineering approach emphasizes explicit trust boundaries, durable processing, database-enforced security, measurable quality, reviewable changes, and production-oriented failure handling.

• Backend: C#, .NET, ASP.NET Core, REST APIs, background services
• Enterprise systems: durable workflows, transactional boundaries, lifecycle/state management, integrations
• Databases: SQL Server, PostgreSQL, pgvector, Row-Level Security, query and schema design
• Architecture: API/Worker separation, least privilege, multi-tenant systems, service trust boundaries
• Security: JWT, durable authorization, tenant isolation, secure document ingestion, negative security testing
• Observability: OpenTelemetry, structured logging, correlation, metrics, health checks, SLO-oriented monitoring
• AI engineering: RAG, semantic retrieval, grounded answers, provider abstraction, evaluation and citation gates
• Infrastructure: Docker Compose, Redis, CI/CD, GitHub Actions, CodeQL, dependency automation

The portfolio is intentionally centered on a small number of reviewable projects rather than repository count. Forks used for upstream contributions are secondary to the projects above.

A local-first enterprise document platform built around ASP.NET Core, FastAPI, PostgreSQL, pgvector, Redis, Docker Compose, semantic retrieval, and grounded AI answers.

Selected engineering capabilities:

• durable document ingestion with transactional job creation, bounded retries, recovery, and PostgreSQL FOR UPDATE SKIP LOCKED claiming;
• JWT authentication with durable tenant membership and immediate authorization revocation;
• forced PostgreSQL Row-Level Security with direct cross-tenant negative tests;
• separated public API, platform-management, and privileged Worker database identities;
• safe TXT/PDF/DOCX ingestion with bounded parsing, OOXML validation, spoofed-file rejection, and explicit OCR-required outcomes;
• persistent pgvector semantic retrieval with reproducible Precision@K, Recall@K, and MRR evaluation;
• provider-neutral grounded-answer generation with mandatory citations and insufficient-evidence handling;
• append-only tenant audit storage, OpenTelemetry traces/metrics, correlation propagation, and operational observability;
• independent CI coverage for application tests, PostgreSQL integration, document formats, retrieval, grounding, Dependency Review, and CodeQL.

Repository: enterprise-ai-document-assistant

Focused contributions merged into established .NET projects:

• dotnet/aspnetcore #67481 — clarified ActionLink URL-generation documentation behavior.
• dotnet/docs #54567 — documented sizeof behavior for enum types in the C# language reference.
• dotnet/docs #54559 — corrected ASP.NET workload documentation in the .NET microservices guidance.

A .NET foundation for provider-independent AI contracts with a deterministic provider, runnable console sample, tests, and CI.

An exploratory concurrency project focused on randomized locking, contention, fairness, reproducible testing, and careful treatment of algorithmic guarantees.

An archived computer-vision study retained with explicit attribution, reproducibility boundaries, and documented limitations.

• Prefer durable state over process-memory assumptions.
• Enforce critical authorization boundaries at more than one layer.
• Treat external inputs and retrieved AI context as untrusted data.
• Make failure modes explicit and observable.
• Use negative tests to validate security boundaries.
• Measure retrieval and answer quality instead of relying on selected demos.
• Keep documentation aligned with implemented behavior.

I am currently deepening work around:

• production-grade multi-tenant identity and authorization;
• enterprise observability, audit integrity, retention, SLOs, and operational runbooks;
• multilingual and adversarial retrieval/answer evaluation;
• secure document-processing boundaries including OCR and complex layouts;
• scalable .NET backend architecture and enterprise data workflows.

• GitHub: @mahdiaghtaee
• For repository-specific technical collaboration, use the relevant project's Issues or Pull Requests so design decisions and evidence remain reviewable.

No additional public contact channel is listed here unless it can be verified and intentionally maintained.