Senior Full Stack Engineer | LLM Systems | Microservices Architecture | Production AI
7+ Years Building Scalable Systems That Ship
I'm a full-stack engineer who's spent the last 7 years building production-grade systems in fintech, logistics, and AI. I specialize in shipping complex features end-to-end—from AWS infrastructure to payment integrations to AI-powered workflows.
- 🔭 Currently Focused On: AI-powered business software, LLM orchestration, production LLM systems
- 💼 Expertise: NestJS microservices, React/TypeScript frontends, AWS infrastructure, LangGraph/LangChain, payment systems
- 🌱 Always Learning: System design at scale, autonomous AI agents, cost optimization
- 💡 Specialized In: Real production problems—not theory. Payment processing, compliance automation, AI workflows
- 🎓 Problem Solver: Battle-tested across fintech (Bank of America, Plaid, Stripe), logistics (50K+ parcels/day), and AI systems (500K+ documents)
- Architected payment processing systems handling millions of daily transactions
- Integrated Bank of America CashPro (RTP/ACH/Wire), Plaid, and Stripe
- Webhook-driven reconciliation, micro-deposit verification, transaction status tracking
- Built entire payment infrastructure on AWS (ECS, RDS, Secrets Manager, CDK)
- Built LangGraph-based agents for document processing and compliance verification
- Implemented RAG pipelines with vector embeddings and semantic search
- Multi-step LLM orchestration with branching logic, retries, and observability
- Handles 500K+ documents with confidence scoring and human escalation
- Designed and shipped greenfield NestJS microservices (complete AWS stack)
- Real-time data pipelines with SQS, RabbitMQ, and Kafka
- Database migrations, schema evolution, audit logging
- Reduced infrastructure costs from $3K+ to under $1,600/month
- Reduced report generation from 30 minutes to under 10 seconds via Kafka denormalization
- 98% performance improvement for batch operations using Kubernetes ephemeral containers
- Implemented Redis caching, database indexing, query optimization
- Bank of America CashPro API integration (RTP, ACH, Wire transfers)
- Plaid API for bank account linking and micro-deposit verification
- Stripe payment integration with idempotency and retry logic
- Webhook handling with signature verification and dead-letter queues
Production system processing 500K+ documents with AI-powered classification
- Built LangGraph-based agent that autonomously processes driver compliance documents
- Implemented vector embeddings and semantic search for document sections
- Multi-step reasoning: extract metadata → classify type → verify compliance → escalate exceptions
- Full traceability logging for audit purposes
- Confidence scoring with human escalation for low-confidence decisions
- Impact: 60% reduction in manual triage work, 100% audit trail compliance
Tech Stack: LangChain, LangGraph, NestJS, PostgreSQL, AWS
Production payment system handling RTP, ACH, and Wire transfers
- End-to-end integration with Bank of America CashPro API
- Real-time transaction status tracking via webhooks
- Micro-deposit verification for account linking
- Webhook-driven reconciliation with dead-letter queues
- Fallback logic and error handling for failed transactions
- Impact: Enabled financial operations for entire platform, zero transaction loss
Tech Stack: NestJS, TypeScript, PostgreSQL, AWS ECS, Bank of America APIs
Complete infrastructure overhaul reducing costs by 70%
- Architected two greenfield NestJS microservices (Client & Broker Services)
- AWS infrastructure as code (CDK/CloudFormation) - deployable via single CLI command
- Real-time sync pipeline between legacy monolith and v2 microservices
- FMCSA DOT compliance checks via EventBridge cron jobs
- MikroORM subscriber-driven audit logging across all entities
- Impact: Monthly infrastructure cost reduced from $3K+ to under $1,600
Tech Stack: NestJS, AWS (ECS, RDS, CDK, EventBridge), Docker, PostgreSQL
Performance improvement from 30 minutes to 10 seconds
- Proposed and architected Kafka-driven denormalization pipeline
- Pre-aggregated parcel/rider/order data into optimized collections
- Eliminated expensive MongoDB lookups
- Impact: 98% performance improvement (30 mins → 10 secs), handled 500K+ parcels/day
Tech Stack: Kafka, MongoDB, NodeJS, Kubernetes
Decomposed monolithic Angular app into independent micro-apps
- Architected Svelte-based micro-frontend shell
- Isolated modules: parcel upload, delivery tracking, reporting, admin
- Shared authentication and state management
- Independent deployment and scaling for each module
Tech Stack: Svelte, NestJS, Docker, Kubernetes
Senior Backend Engineer
- Architected and shipped greenfield microservices on AWS
- Integrated complex payment systems (BOA, Plaid, Stripe, Modern Treasury)
- Built LLM agents and RAG pipelines for compliance automation
- Led team mentorship and cross-functional coordination
Senior Full Stack Engineer
- Designed and shipped Kafka-driven reporting system (98% faster)
- Architected microservices migration from LoopBack 3 to NestJS
- Hired and trained 5 junior engineers
- Proposed and built Svelte micro-frontend architecture
Senior Full Stack Developer
- Integrated Stripe + Apple Pay + Google Pay with wallet debit logic
- Implemented role-based access control for multi-user accounts (200K users)
- Built RabbitMQ message queuing infrastructure
Senior Full Stack Developer
- Built JSON schema → NestJS code generation engine
- Auto-generated GraphQL resolvers, entities, migrations, and endpoints
- Designed multi-language support with NextJS + Formik
Full Stack Developer
- Shipped production logistics MVP in 3 months (50K+ parcels/day)
- Optimized reporting with Kafka denormalization (30 mins → 10 secs)
- Built reusable Angular components and data recovery scripts
1. Ship Working Code, Not Demos
→ Production ready from day one
→ Observability, error handling, monitoring built-in
→ Real-world edge cases addressed
2. Architecture Matters
→ Clean separation of concerns
→ Scalable design from the start
→ Cost-conscious infrastructure decisions
3. Own the Full Stack
→ Database schema to UI deployment
→ Understanding failure modes
→ Mentoring others along the way
4. AI as a System Component
→ Treat AI like any backend service
→ Observability, cost, error handling are paramount
→ Confidence scoring and human escalation patterns
5. Code is for Humans
→ Clear naming and structure
→ Comprehensive error messages
→ Documentation where it matters
✅ 7+ Years Production Experience - Not theory, real systems at scale ✅ Full-Stack Ownership - Architecture → Frontend → Infrastructure → Deployment ✅ Problem Solving - Payment systems, compliance automation, performance optimization ✅ Architecture Leadership - Design scalable systems that don't explode costs ✅ Team Mentorship - Train junior engineers, establish best practices ✅ AI Systems - Production LLM workflows, not chatbot experiments
I'm available for:
🔹 Backend Development - NestJS microservices and scalable APIs 🔹 Full-Stack Projects - End-to-end feature ownership 🔹 AI Systems - LLM orchestration, RAG pipelines, agent workflows 🔹 Architecture Design - Microservices, payment systems, compliance automation 🔹 Performance Optimization - Database tuning, caching strategies, async patterns 🔹 Team Leadership - Code review, mentorship, technical direction
- 💼 Upwork: Hire Me on Upwork
- 💬 LinkedIn: Connect on LinkedIn
- 📧 Email: [email protected]
- ☎️ Phone: +92 304 7216339
"Production systems aren't impressive because they're complex—they're impressive because they work reliably, scale gracefully, and fail gracefully when they do fail."
Building software is about solving real problems for real people. I care about shipping working solutions, not collecting likes on cool tech.
⭐️ From Ehsan - Building production systems, one service at a time.