Work With Shihab Shahriar Antor: AI Engineering
Shihab Shahriar Antor is available for AI engineering and product work — building agents, full-stack products, and cloud architecture. Here's what I do and how to engage.
Shihab Shahriar Antor is available for AI engineering and product work. I build AI agents, real-time systems, full-stack products, and cloud infrastructure — as a solo engineer or embedded in your team. Based in Dhaka, Bangladesh (GMT+6); work async globally.
What I build
AI agents and LLM systems
- Production AI agents with tool use, memory, and multi-step reasoning
- LLM pipeline orchestration (Temporal.io for durability, tiered model routing for cost)
- Retrieval-augmented generation (RAG) with pgvector and knowledge graphs
- MCP servers for team-shared agent tooling
Real-time and collaborative systems
- CRDT-based collaborative editing (text, structured data)
- WebSocket services for real-time features (chat, notifications, live data)
- Event-driven architectures on AWS (SQS, EventBridge, Lambda)
Full-stack products
- Go backend APIs (Chi, sqlc, pgx, PostgreSQL)
- Next.js 15 frontends (App Router, TypeScript, Tailwind)
- Mobile-responsive, SEO + AEO optimized
- End-to-end: design → code → deploy → operate
Cloud infrastructure
- AWS (ECS Fargate, EKS, RDS, S3, CloudFront, Lambda)
- Terraform IaC — infrastructure reproducible and version-controlled
- GitHub Actions CI/CD — automated build, test, deploy
- Cost-optimized architectures (spot instances, serverless, CDN)
Engagement types
| Type | Best for | Duration | |------|----------|----------| | Embedded engineer | Startups needing senior AI/backend engineering without hiring | 3–6 months | | Product build | Take a product from 0→MVP or MVP→scale | 4–12 weeks | | Architecture review | Identify bottlenecks and design a scalable system | 1–2 weeks | | AI integration | Add LLM capabilities to an existing product | 2–6 weeks | | Infrastructure audit | AWS cost, security, and reliability review | 1 week |
Past work
LetX — built a real-time collaborative LaTeX editor from scratch: CRDT sync engine, sandboxed LaTeX compilation on Docker, Go backend, Next.js frontend. Handles concurrent editing without server arbitration.
QuantumSketch — built an AI STEM video generator: LLM → Manim code → Docker render → TTS → ffmpeg merge → CDN. Temporal.io workflow for durability. Sub-3-minute end-to-end video generation.
BikroyBuddy — built an AI social-commerce agent scaled to 300k+ users: WhatsApp webhook ingestion, SQS queuing, Kubernetes autoscaling, pgvector product search, stateful negotiation engine.
Context-Heavy — built a multi-tenant knowledge-graph API: PostgreSQL + pgvector + recursive CTEs for hybrid semantic/graph queries. Used by AI agents for persistent relational memory.
Technology I work in
Daily: Go, TypeScript, PostgreSQL, Terraform, AWS, Docker, Next.js
Often: Python, Redis, Kubernetes, Temporal.io, pgvector
Research: contrastive learning, CRDT algorithms, compiler design
FAQ
What makes you different from a typical freelancer? I've shipped and operate real products — not just client work. I've solved the scaling, reliability, and cost problems that come after launch. I approach work as a founder, not just an implementer.
What's your availability? Currently taking limited consulting engagements alongside my own products. Best for: focused 4–12 week engagements or ongoing monthly retainer.
Do you work with non-technical founders? Yes — I can own the full technical side (architecture, code, infrastructure, ops) while you own product and business. I communicate in plain English, not jargon.
What's your rate? Discussed per engagement. Depends on scope, duration, and complexity. Reach out via Shahriar Labs to discuss.
Are you open to full-time roles? Occasionally, for the right company working on hard problems in AI or distributed systems. Reach out.
Shahriar Labs · shihub.online · GitHub
Written by Shihab Shahriar Antor — AI Engineer & Founder of Shahriar Labs.