Decebal Dobrica
AI Engineer, Architect & Leader
From GenAI strategy to production RAG systems
Other Expertise
About
With 5+ years of AI/ML experience across fintech, SaaS, and enterprise platforms, I help organizations navigate the GenAI revolution. From strategic adoption to production RAG systems, my approach bridges the gap between executive vision and engineering reality, turning AI skeptics into advocates.
At CyberSpark Systems, I led a GenAI transformation that achieved 40% sprint acceleration by reframing AI as a co-pilot rather than a replacement. I created daily intelligence dashboards, rolled out targeted AI tools, and fundamentally changed how the team approached development.
My AI engineering expertise spans the full stack: from building production RAG systems with AnythingLLM and Groq, to integrating LLMs via the Vercel AI SDK, to pioneering AI-assisted development workflows that amplify developer productivity.
Expertise
GenAI Strategy & Adoption
Leading organizations through GenAI adoption, transforming skeptical teams into AI advocates. Achieved 40% sprint acceleration by bridging the gap between engineering and executive vision.
RAG Architecture
Building production RAG systems with vector databases, embeddings, and intelligent retrieval. Expertise in AnythingLLM, LanceDB, and custom RAG pipelines for personalized AI experiences.
LLM Integration
Integrating LLMs into production applications with Groq, OpenAI, Anthropic, and local models. Optimizing for latency, cost, and quality with streaming responses and intelligent caching.
AI-Assisted Development
Pioneering AI-assisted development workflows with Claude Code, GitHub Copilot, and custom tooling. Building systems where AI amplifies developer productivity without replacing human judgment.
Relevant Case Studies
View allRelated Articles
View allRust & AI Weekly #10: what runs before your code does
Today's issue: a crate with 245 million downloads shipped a build script that phoned home for 86 minutes, the Rust tool that would have sandboxed it has been sitting there since 2023, and the ecosystem finally put money on the maintainers whose credentials are the attack surface. Each entry carries maintenance, latest-release, and adoption signals.
Rust & AI Weekly #9: where the boundary is drawn
Today's issue: a Rust sandbox whose manual names AI/ML workloads as the thing to contain, the maintainer of rustls publishes a memory-safe SSH server after 20 months and says plainly that it is not ready, and rama's release-train promise from issue #4 comes due. Each entry carries maintenance, latest-release, and adoption signals.
Rust & AI Weekly #8: make it unrepresentable
Today's issue: fearless_simd takes unsafe out of SIMD and books 1.0 for September, a zero-dependency crate makes quantized-tensor mismatches fail loudly instead of quietly, and the Rust SVG toolchain Parcel already ships grows an SVGR replacement. Each entry carries maintenance, latest-release, and adoption signals.
Rust & AI Weekly #7: publish the audit
Today's issue: webrtc-rs ships its Sans-I/O rewrite and grades its own roadmap, FalkorDB moves 80,000 lines of graph engine to Rust for GraphRAG, and a btop for your Claude Code context window. Each entry carries maintenance, latest-release, and adoption signals.
Rust & AI Weekly #6: the foundations move
Today's issue: Syn ships its first major in three years, SeaORM 2.0 goes stable after 43 release candidates, and a Claude SDK that can't panic. Each entry carries maintenance, latest-release, and adoption signals.
Rust & AI Weekly #5: the officials arrive
Today's issue: BullMQ ships an official Rust port with benchmark receipts, xAI open-sources 845,000 lines of coding agent, and Tokio tries on a full-stack web framework. Each entry carries maintenance, latest-release, and adoption signals.
Other Expertise
Ready to Work Together?
Let's discuss how I can help with your ai engineer, architect & leader needs.
Discuss AI Strategy