Dennis S.

Software engineer with 10 years of experience across full-stack applications, backend systems and native Windows software. Specializing in production AI systems, C++, Python. Combining LLM application development with application architecture, performance engineering and end-to-end software delivery. Top Rated Plus on Upwork with 100% Job Success. Development of AI-powered products from architecture and model selection through data integration, deployment and evaluation. Focus on reliable systems with measurable quality, controlled operating costs and application-enforced security. AI Agents & LLM Applications AI applications using OpenAI, Claude and open-weight models. Agentic workflows with tool calling, structured outputs, persistent conversation state and MCP integration. Workflow orchestration with bounded execution, failure recovery and human approval for consequential actions. Model and architecture selection based on accuracy, latency, privacy, reliability and cost. Integration of AI capabilities into existing products and business systems. RAG & Agentic Search Retrieval-augmented generation connecting models to documents, databases and live information. Document ingestion, parsing, chunking, embeddings and vector indexing. Hybrid search, metadata filtering and reranking to improve retrieval relevance. Source-grounded responses with citations and access-controlled retrieval. Agent-directed search with iterative information gathering, freshness checks and explicit handling of insufficient evidence. Evaluation, Guardrails & Model Customization Evaluation datasets and repeatable tests for response quality, retrieval accuracy, tool execution and regression detection. Measurement of task completion, latency, token usage and cost. Application-enforced permissions, schema validation, sensitive-data protection and layered prompt-injection defenses. Supervised fine-tuning and parameter-efficient adaptation, including dataset preparation, model comparison and validation against baseline performance. Conversational & Voice AI Real-time voice systems using Telnyx, Retell AI and direct OpenAI integration. SIP connectivity, WebRTC, WebSocket audio streaming, speech recognition and speech synthesis. Conversation state, interruption handling, backend tool execution and transcript persistence. Signed webhook verification, duplicate-event handling and integration with business workflows. AI Deployment & Production Operations Deployment and maintenance of AI applications using Docker, AWS and Cloudflare. Model and prompt versioning, automated testing, production monitoring and controlled releases. Tracking response quality, latency, token usage and operating costs. Reliable API integrations with retries, fallback handling and rollback strategies. Full-Stack & Performance Engineering React and TypeScript interfaces with Python, FastAPI and Node.js services. Database architecture, authentication, role-based access, Stripe subscriptions and usage controls. Dashboards, administration interfaces and secure content delivery. Native Windows development in C++ and Win32, including multithreading, memory management and performance profiling. Modernization and integration within established codebases. Custom Trading Applications supporting Rithmic, IBKR, Alpaca, Charles Schwab brokers data and order routing. Project scope ranges from focused AI integrations to complete application development, covering architecture, implementation, testing and deployment.