AI development and AI agents
AI work here ranges from a single feature to a full agentic system. I build agents that plan, call tools, and finish multi-step jobs — with LangChain, LangGraph, CrewAI, and AutoGen when they help. That includes tool-calling, multi-agent orchestration, structured reasoning loops, memory, and human-in-the-loop guardrails.
If you need to build an agent, wire OpenAI, Anthropic, or Gemini into a product, or ship an autonomous workflow, this is core work for me. I currently lead AI on a production agentic SaaS platform.
LLM apps, RAG systems, and chatbots
LLM product work means turning models into features you can operate. With RAG, I connect your documents and databases so answers stay grounded and cited — less hallucination, more trust. That uses vector stores (Pinecone, Weaviate, Qdrant, ChromaDB, pgvector), embeddings, hybrid search, reranking, and prompt evals.
Common builds: a support chatbot, an internal knowledge assistant, document Q&A, or a voice agent. Chatbots can live on your site, WhatsApp, or inside your app.
Generative AI (image, voice, and video)
Generative features add media capabilities: text-to-image (ComfyUI, Replicate, FAL.ai), speech-to-text and TTS, real-time transcription, OCR, face swap, and video/audio pipelines. I also build the GPU inference paths and queues (RunPod, cloud GPUs) so these hold up under real traffic.
Custom software and backends
Custom software means software shaped around your business logic — internal tools, admin dashboards, portals, and data pipelines. On the backend I design APIs and services with FastAPI, Django, and Node.js, using PostgreSQL, Redis, and MongoDB. That covers REST and GraphQL, event-driven patterns (Kafka, Celery), auth (JWT/OAuth 2.0, RBAC), and performance work.
Web apps, websites, and SaaS MVPs
Web app and website work uses Next.js, React, and TypeScript up front with Python or Node behind. For founders, SaaS MVP work gets a launchable slice — auth, Stripe/PayPal billing, multi-tenancy, usage/credits, and AI features where they help. Fits building a SaaS, shipping an MVP, or adding AI to an existing app.
Business automation, API integration, and cloud
Automation replaces repetitive manual work with schedules, webhooks, queues, and event pipelines. I also handle API integration with Stripe, CRMs (HubSpot, Salesforce), and internal services, plus notifications and reporting. Delivery usually includes Docker, Kubernetes, CI/CD (GitHub Actions), AWS/GCP, and monitoring with Sentry and similar tools.