Mykola Melnyk

- Architected and deployed a production-grade AI assistant for a Fleet Management System using LangGraph, enabling complex agentic workflows for logistics operations. - Engineered multi-intent handling with NeMo Guardrails, ensuring the system adhered to strict corporate security protocols and maintained a near-zero hallucination rate. - Real-Time Integration: Built high-performance integration to fetch and process live operational data, allowing for dynamic fleet adjustments via natural language.

PDF Redaction - AI powered tool for detect and remove sensitive information from the PDF documents: - Online tool - API - On-Premise

I'm a creator of https://pdf-redaction.com - AI-powered pdf redaction online tool (ScaleDP/Next.js/React/FastAPI/LLM/OCR/NLP).

I'm developing Lead Generation solution using LangChain/LangGraph.

- Led developing Visual NLP (Spark OCR). Spark OCR is commercial extension of Spark NLP for optical character recognition from images, scanned PDF documents, Microsoft DOCX and DICOM files. - Designed architecture and implemented big data solutions using Spark OCR and Spark NLP. - Collaborated with stakeholders to define technical requirements and deliver high-impact AI solutions.

Led developing Data Science infrastructure based on Spark. Technologies: Big Data, ETL, Scala, Python, Spark, Spark Structured Streaming, Spark MLlib, Databricks, AWS, DynamoDB, Kinesis, Kafka, Graphite, Grafana, Jenkins.

I lead developing and maintenaning CloudML - declarative machine learning infrastructure. Technologies: Big Data, Python, Flask, AngularJS, Scikit-Learn, XGBoost, Spark, Hadoop, PIG, Celery, AWS( S3, EC2, EMR, DynamoDB), Spark, Graphite, Grafana, Logstash, Kibana.

Lead developing CloudML - declarative machine learning infrastructure project. Technologies: Python, Flask, Scikit-Learn, Numpy, Scipy, Hadoop, PIG, Celery, AngularJS, PostgreSQL, DynamoDB. CloudML functionality: - defining datasets and ML models - managing/versioning it - train/test ML models - visualisation models - deploying and serving models as high availability and horizontal scalable service

Developed Django based task management services for integrate enterprise clients to oDesk. Technologies: Python, Django, Flask, AngularJS, AWS, Celery, RabbitMQ, Redis

Developed web sites using Python, Django, Javascript. Technologies: Python, Django, Satchmo, Pinax, Javascript, JQuery, Nose, Twill, tddspry

Core expertise of Ltd Argus includes development, implementation, support of complex software products, software and hardware complexes for the automation of financial activities of credit and financial institutions and enterprises. While Argus Ltd, I developed and supported software for payment terminals and ATM’s. I successfully implemented new features in time. I adapted software for new hardware.

Projects: ICQ Proxy is proxy server with encryption messaging. Project was implemented in time and consumers started to use it. TRNG is hardware true random number generator. Hardware is Atmel ARM7. Interface is USB. I designed hardware and developed software. Developed drivers for windows and Linux. UkrCos is a system for managing private keys based on smart cards. I developed hardware and some part of software: smart card software, custom drivers for windows and Linux, cross-platform API. Implemented so

Projects: Automatic control system for lifting machines in the mines. Power of main engine is 5 megawatts. I collected requirements, designed and developed hardware and software. I implemented control algorithms, alarm system, visualisation. I lead two developers in team. System was certified and performed commissioning on “Comsomolec Donbasa” coal mine. Automation control system for heavy soda production. I collected requirements, designed and developed part related main calciner. Hardware: Schneider indu