Anton S.

Project overview:

Built the Agentic AI runtime for a global consumer AI chat platform with 900K+ paid subscribers and $1M+ daily net revenue. The goal was to evolve the product from a chat wrapper into a true agentic platform — multi-step reasoning, real-time tool use, and consent-gated actions — shipped as a zero-downtime additive layer on a live production system.

Responsibilities:

• Architected the full agentic runtime: context assembly layer, multi-turn agent loop with live tool calls, streaming responses, error recovery, and token budget management across 12 global regions.
• Designed the MCP-first connector architecture with zero-central-edits isolation — each connector is a self-contained folder discovered by codegen, so parallel work across dozens never causes merge conflicts.
• Built the LLM evaluation and prompt-optimisation framework: 5-dimension LLM-as-judge scoring, failure-driven candidate generation, and automated weekly prompt improvement via Langfuse.

Achievements:

• Shipped a production-ready agentic MVP across 5 phases with zero downtime, enabling the product to compete directly with ChatGPT and Manus. • Headed the Connectors department (team of 11) and designed its entire operational process from scratch — an 8-stage Auth → Backend → Frontend → QA → Release pipeline that ships 5–8 connectors to production per day without interrupting ongoing development. Embedded Claude-powered AI agents at every stage so engineers supervise and approve rather than execute, enabling the platform to offer 4,000 integrations where users read emails, search docs, create tasks, and send messages — all from a single AI chat. • Drove P0 cost optimisation on daily inference spend — model routing, context compaction, prompt A/B testing — protecting margins while guarding output quality via LLM-as-judge eval gating.

Technology stack:

Project overview:

Designed and delivered multiple AI-driven product features and strategies across consumer SaaS — spanning a media AI expansion, an AI-powered job application platform, and first-line support automation. Each project combined product thinking with hands-on engineering to produce measurable business outcomes.

Responsibilities:

• Verified all model pricing and availability independently — correcting estimates up to 50x too high, preventing broken launches, and producing board-ready strategy documents with ICE-scored prioritisation.
• Built multi-agent architectures (LangGraph, CrewAI) for job matching, interview coaching, and support triage — reducing manual workload by 60–80% with 24/7 autonomous operation.
• Engineered cost-optimised hybrid AI workflows with model routing between cloud and local LLMs (DeepSeek, Llama 3), cutting inference costs by ~40% without accuracy regression.

Achievements:

• Media AI: built the business case for 8 fal.ai photo/video features competing with $200M+ revenue standalone products — projecting 7–37x ROI in Month 1 at $200–$1,000/mo infrastructure cost, with +1–2% subscription conversion lift. • Job Platform: reduced time-to-apply from 2–3 hours to under 5 minutes via an end-to-end AI pipeline (job scraping, CV analysis, role-tailored CV/cover letter generation, auto-apply, AI interview coach) — achieving 3x higher interview callback rates. • Support Automation: delivered 60–80% reduction in Tier-1 support ticket volume with positive ROI within 90 days, using a RAG system over 280+ knowledge base articles with automated hallucination detection.

Technology stack: