Senior Full-Stack Software Engineer with 12+ years shipping distributed systems, Python backends, and AI agents for fintech, banking, and enterprise platforms - currently at Apex Fintech Solutions.
Selected work
Surfaces I've owned end-to-end across fintech, banking, vehicle telemetry, and large-scale content platforms.
Designed multi-step LangChain/LangGraph agents that investigate transaction alerts, review customer profiles, and produce audit-ready narrative reports for bank compliance teams - directly cutting manual analyst workload on BSA/AML reviews.
Built resilient Python ingestion services integrating 10+ financial data providers via REST and webhook APIs - Pydantic validation, idempotent processing, retries, and structured error handling for transaction data, KYC records, and AML alert feeds.
Built async Python FastAPI microservices ingesting hundreds of millions of daily vehicle sensor events. Pydantic schema validation, asyncio I/O, and horizontal scaling on AWS EKS for large-scale telemetry analytics.
Developed LLM-powered internal tools over structured vehicle telemetry - natural-language querying of sensor metrics and automated anomaly insight generation, reducing time engineers spent interpreting trend data by ~30%.
Designed Kafka event-driven data pipelines with idempotent consumers, exactly-once semantics, and dead-letter routing - enabling millisecond-latency propagation of critical vehicle alerts across distributed backend services.
Built Python FastAPI/Flask backend services for enterprise banking data APIs processing millions of daily transactions - multi-tenant PostgreSQL schemas, Pydantic validation, JWT auth, RBAC, and OpenAPI docs.
Developed LLM-powered compliance document analysis with LangChain and OpenAI - RAG pipelines over FAISS for regulatory document retrieval, automated summarization, and semantic search, reducing manual review workload ~35%.
Built Kafka event-driven pipelines for real-time transaction propagation - idempotent consumers, exactly-once semantics, dead-letter routing, and millisecond-latency fraud detection event delivery to downstream Python services.
Built Python Django/Flask backend services for a manufacturing operations platform serving 50,000+ daily users - REST APIs, multi-domain PostgreSQL schemas, JWT auth, and integrations with 10+ data providers and logistics APIs.
Designed and tuned PostgreSQL schemas across 7 years of iterative growth - normalization, composite indexing, window-function reporting queries, and pgBouncer pooling, cutting report generation from 5 minutes to 30 seconds on multi-million-row datasets.
Built distributed async task pipelines using Celery and RabbitMQ - retries with exponential backoff, dead-letter queues, and idempotent handlers for ERP synchronization, batch reporting, and logistics integrations.
Developed ML workflows with Pandas, NumPy, and scikit-learn for demand forecasting and inventory optimization - served model outputs through REST endpoints powering real-time predictive insights in operational dashboards.
Built ETL data pipelines with Python and Apache Airflow - multi-source ingestion from REST APIs and database change events into PostgreSQL and Elasticsearch with validation, deduplication, and automated quality alerting.
Built Python backend services and REST APIs for content management platforms serving 10M+ monthly users - owned features end-to-end from PostgreSQL schema design through production deployment at global scale.
Built BigQuery and Hadoop MapReduce data pipelines for large-scale analytics - ETL workflows with validation and error handling, plus Memcached caching that cut MySQL query load 60% and retrieval latency from 300ms to 50ms.
Implemented early LLM and semantic-search experiments using TensorFlow and Elasticsearch - trained embedding models on structured content and indexed vectors for semantic retrieval, prototyping natural-language content discovery features.
Career
Shipping Python backends and LangChain/LangGraph compliance agents for BSA/AML automation.
FastAPI telemetry services on EKS, Kafka pipelines with exactly-once semantics, and LLM tools over sensor data.
Multi-tenant banking APIs, RAG over compliance docs with FAISS, and real-time fraud detection on Kafka.
Django/Flask manufacturing platform for 50K+ daily users, PostgreSQL tuning, Celery pipelines, ML forecasting, Airflow ETL.
Backend services for content platforms at 10M+ MAU; BigQuery/Hadoop pipelines; early embedding-based semantic search.