Ali A.

I add AI to products that already work, and I keep them running. Senior full-stack engineer (Ruby on Rails, Python, Elixir/Phoenix) with 8+ years of production experience, now focused on AI integration, chatbots and RAG for SaaS teams. 𝐑𝐄𝐒𝐔𝐋𝐓𝐒 𝐈 𝐂𝐀𝐍 𝐒𝐇𝐎𝐖 - GPT-4 ticket classification for a global travel technology company: 10K+ tickets a month, 95% accuracy, about $200K saved a year, with confidence scoring and human-review fallback - AI analytics platform: GPT-4 report generation and natural-language data querying for 50K+ monthly users - Semantic search on OpenAI embeddings and pgvector: search satisfaction up from 45% to 89%, "can't find it" support tickets down 70% - Background processing at scale: 500K+ async jobs a day with Sidekiq and Redis - Rails 5.2 to 7.1 migration with a 30% performance gain - [ADD: your best AI chatbot project, with one number] 𝐖𝐇𝐀𝐓 𝐈 𝐁𝐔𝐈𝐋𝐃 - AI chatbots and assistants (OpenAI, Claude) that answer from your own documents and data (RAG) - LLM features inside existing Rails, Python or Node apps: classification, summarisation, search, report generation - AI agents and workflow automation with tool calling, structured outputs and human-in-the-loop fallbacks - Rails upgrades, performance tuning and long-term backend ownership - Real-time systems in Elixir/Phoenix (100K+ concurrent connections) 𝐇𝐎𝐖 𝐈 𝐖𝐎𝐑𝐊 - A short scoping call first, then a fixed-price plan or a clear hourly estimate - Weekly demos and written updates, working inside your Slack, Jira and GitHub - Production habits: tests, monitoring, LLM cost tracking, and fallbacks for when the model is wrong 𝐒𝐓𝐀𝐂𝐊 AI: OpenAI GPT-4, Claude API, LangChain, pgvector, Pinecone, Elasticsearch Backend: Ruby on Rails, Python, Elixir/Phoenix, Node.js, GraphQL, gRPC, PostgreSQL, Redis, Sidekiq Cloud and DevOps: AWS, Docker, Kubernetes, GitHub Actions, Datadog, New Relic Frontend: React, TypeScript, Angular 𝐒𝐄𝐋𝐄𝐂𝐓𝐄𝐃 𝐏𝐑𝐎𝐉𝐄𝐂𝐓𝐒 - Milaap: backend lead for a crowdfunding platform serving millions of users (Rails, PostgreSQL, Redis, AWS; clustering, replication, load balancing) - AppWorkCo: led a 30+ engineer team on a multi-tenant property-management SaaS (Elixir/Phoenix, React, RabbitMQ, Docker, AWS) - DMARC Reporting Suite: multi-tenant compliance platform (Rails, Sidekiq, PostgreSQL, AWS Lambda) Tell me what you want the AI to do and what stack you run today, and I'll reply with a scoped plan within 24 hours.