Let's Build Something
Amazing Together
15+ years building software products and leading engineering teams, 2+ years deep in AI. Now helping companies ship production agentic AI apps and re-imagine their product vision.
What I Bring to the Table
Agentic AI Apps
Production-ready AI agents — orchestration, evals, observability, and computer vision. From rapid prototype to scalable system. LangGraph, LLMs, RAG.
Full-Stack Development
Native iOS/Android, React, TypeScript, real-time streaming. Modernize interfaces across mobile and web. Built apps with 6K+ paying subscribers.
Product & Technical Leadership
End-to-end product development, team building, and engineering process. Founded startups, led cross-functional teams, shipped products from zero to thousands of users. Fundraising and go-to-market strategy.
Experience
Independent Consultant
Fairmarkit
Shipping two production AI systems for a B2B procurement marketplace: Bid Pulse (agentic RFQ monitoring) and Fair Platform (an internal Lovable that lets product managers and sales vibe-code prototypes straight to clients — securely, with production-grade infrastructure).
Bid Pulse: built end-to-end agentic system that discovers RFQs, matches them against 2.5M+ supplier products via three-tier engine (MPN exact match → keyword scoring → pgvector semantic search) with dual-evaluator LLM verification
Fair Platform: AI-assisted IDE with live preview, git-based template scaffolding, one-click publishing, and per-app Kubernetes isolation — so PMs and sales can ship to clients without engineering capacity
Security hardening for Fair Platform: Squid egress proxy with FQDN allow-list, Falco runtime threat detection with tiered alerts, NetworkPolicy blocking metadata access, Okta OIDC with JIT provisioning
Stack: FastAPI, Next.js, React 19, PostgreSQL/pgvector, Kubernetes (DOKS), Terraform/OpenTofu, ArgoCD, OpenAI, Anthropic Claude Opus, Langfuse
Independent Consultant
Zumbach Electronic
Modernizing the UX of a Swiss industrial measurement system whose interfaces had been frozen in the 90s. Conducted user research with factory floor operators, then designed and built a new web-based monitoring product from scratch — replacing legacy USYS hardware displays.
Real-time live dashboard with 4-channel measurement grid and dual view modes; ZeroMQ pub/sub streaming from device → FastAPI → WebSocket to React
Visual centering guidance for 1/2/3-axis laser configurations with adaptive geometry and interactive flip view
Touch-screen-optimized tolerance configuration and product-recipe editing for factory operators
Multi-arch Docker builds (amd64, arm64, arm/v7 for Raspberry Pi) and a custom monorepo UI component library shipped via tsup + Storybook
Head of Product Development
Winrate.com
Built the entire AI agent stack, call transcript ingestion, evaluation framework, and distributed infrastructure for a B2B sales platform from scratch. Managed a team of four engineers with full ownership of product management, roadmap, and delivery.
Built ICP creation, ICP scoring, account research, research-document, meeting-prep, and general-chat agents on LangGraph with PostgreSQL checkpointing and human-in-the-loop interrupts
Call transcript ingestion across Gong and Fireflies — three-stage pipeline (poller → SQS → processor) with chunking, embedding (OpenAI text-embedding-3-small), and pgvector storage
Evaluation framework with multi-model comparison (Claude-3.5 vs GPT-4) and dataset-driven repetitions, integrated with LangSmith
Reduced wasted Lambda invocations by 96% — from ~30x to 1.1x per job at 15K jobs/day — via Redis Lua token-bucket rate limiting and a backpressure-aware singleton dispatcher
Co-founder, CTO
Enabled Health Inc
AI physical therapy for hospitals. Raised $200K pre-seed, won MA Digital Health Sandbox grant, launched two hospital pilots. Closed in February 2026 after concluding the team could not establish sustainable revenue against long hospital sales cycles; returned remaining capital to investors.
iOS bedside trainer integrating computer vision–based engagement tracking and exercise guidance for hospitalized patients
Voice AI agent in production to communicate with patients, deliver exercise prompts, and collect feedback
Led R&D of computer vision algorithms for body-pose estimation and motion tracking
Managed a mixed remote and on-site team of engineers, UX and motion designers, and data analysts
Conducted hospital staff and patient interviews to validate features
Founder & CEO
Ace Trace
Mobile sports video editor and shot-tracer app for golf, disc golf, and baseball. Built solo and sold in 2025 to an Italian app portfolio company through Flippa, with full upfront payment via escrow.com.
Grew to 6,000 paying subscribers, $15K MRR, ~10,000 monthly installs, 4.4-star App Store rating, #2 on App Store for "golf tracer"
Custom UNet-RCNN neural network detecting 4-pixel objects in video, deployed on-device via iOS On-Demand Resources
Physics-based trajectory fitting with NLopt + Boost/odeint compiled as iOS DyLib; OpenCV video stabilization; visual SLAM camera pose estimation
Hired and managed contractors for UX, ML research, and QA; ran weekly QA testing meetings and extensive A/B tests on retention, pricing, and onboarding
Lead Software Developer
ALM Works
One of the early engineers building Structure for Jira — the flagship enterprise project portfolio plugin. ALM Works grew to 5,000+ customers and was acquired by Tempo Software in 2021. Led a team of four building Structure.Gantt from scratch.
Built Structure.Gantt from initial concept to 5,000+ installations with full architectural ownership; led a team of four engineers
Joined the Labs team (Structure Innovations) — researched new product ideas, ran customer interviews, prototyped MVPs, and shipped Cross-Team PI Planner
Established new development processes, interviewed engineering candidates, and mentored junior developers across seven years
Served as technical escalation point for complex support issues; presented company products at international conferences
Technical Skills
AI & Machine Learning
- Agentic AI systems and multi-agent workflows
- LangGraph/LangChain for AI workflows
- Computer vision (pose estimation, object tracking)
- Voice agents and chatbots using LLMs
- PyTorch
Mobile Development
- Swift, SwiftUI
- Kotlin, Jetpack Compose
- React Native, Flutter
- iOS and Android native development
Backend & Cloud
- Python, Java, Kotlin
- Node.js, TypeScript
- AWS (Lambda, CDK, API Gateway)
- Google Cloud, Firebase
- PostgreSQL, MongoDB
- Docker, REST APIs
- Scalable serverless architecture
Frontend
- React, Redux
- TypeScript
- Real-time streaming interfaces
- Modern responsive design
- UX design
Book a Consultation
Schedule a 1-hour consultation to discuss your project. Whether you need help with AI integration, mobile app development, or product strategy, I'm here to help.
Or reach out directly:
alex@ivkin.dev