Rehan Tariq

Applied AI and Full-Stack Software Engineer with a Bachelor of Science in Computer Science from FAST-NUCES. I specialize in designing and shipping production-grade, agentic AI systems, real-time voice streaming pipelines, and robust Retrieval-Augmented Generation (RAG) architectures.

My engineering philosophy centers on a zero-assumptions approach: using probabilistic LLM reasoning while enforcing strict deterministic backend boundaries (such as Pydantic validation and structured JSON outputs) to guarantee reliability in production. I work extensively across the full stack using TypeScript, Next.js, Node.js, Python, FastAPI, and PostgreSQL.

Key Technical Expertise & Projects:

• Agentic Workflows & Tool Calling: Architected multi-agent orchestration frameworks (using LangGraph) for automated webhook analysis, code review pipelines, and autonomous SRE command systems.

• Real-Time Voice AI: Built high-throughput WebSocket infrastructures integrating persistent local Whisper STT and Coqui TTS workers to eliminate model load latency for bidirectional audio streaming.

• LLMOps & Evaluation: Developed custom asynchronous evaluation harnesses and integrated OpenTelemetry/OpenInference to trace agent decision trees and grade execution trajectories across hundreds of operational scenarios.

• Modern Developer Tooling: Heavily leverage agentic development workflows (Claude Code, Antigravity) to rapidly scaffold, test, and iterate on complex software systems.

I thrive in high-velocity, product-focused startup environments where I can take full ownership of features from architecture through deployment and operational observability.