Programming & Backend
Python, SQL, JavaScript, FastAPI, Django, Flask, REST APIs, SQLAlchemy, Alembic, production backend architecture
About
Novi Pazar, Serbia | +381 62 266996 | resadspahovic94@gmail.com | LinkedIn: Resad Spahovic | GitHub: resadsp | resadspahovic.com
Python Engineer | AI Systems Engineer | Backend, Data & ML Infrastructure
Experienced in building production-ready Python services, ML workflows, LLM/RAG pipelines, and data platforms with measurable impact across multiple industries.
- Birthday: 17 January 1994
- Phone: +381 62 266996
- Location: Novi Pazar, Serbia
- Age: 32
- Degree: Master's Degree
- Website: resadspahovic.com
Focus
AI Systems
Production Python services, ML workflows, LLM/RAG pipelines, and trusted retrieval systems.
Impact
Measured Delivery
Built systems for 50,000+ daily users, cut a nightly pipeline from 4 hours to 45 minutes, and now run my own healthcare software company.
Domains
Cross-Industry Experience
Healthcare, legal tech, agriculture, datacenter operations, analytics, and technical education.
Professional Summary
Python Engineer with 4+ years of experience specializing in AI systems, backend development, data engineering, and ML infrastructure.
Experienced in building production-ready Python services, ML workflows, LLM/RAG pipelines, and data platforms across healthcare, legal tech, agriculture, datacenter operations, and analytics. For the past two years I have designed and built a clinic management system that covers an entire healthcare provider — scheduling, records, laboratory, operating theatre, home visits, stock and finance — and I now run it as my own company, MAAT INNOVATIONS. Alongside it: applications serving 50,000+ daily users and data pipelines rewritten from 4 hours down to 45 minutes.
50,000+
Daily active users served by ML-powered applications.
20
Clinic modules in one system — from booking to profit per department.
4h to 45m
Optimized data processing pipelines for faster operational delivery.
2,400+
Automated checks run before every change to the clinic platform ships.
Core Technical Stack
The site content below follows the same categories and technologies listed in the CV.
AI / Machine Learning
TensorFlow, PyTorch, Scikit-learn, Pandas, NumPy, OpenCV, NLP, Computer Vision, Time-Series Forecasting
LLM / Retrieval
RAG pipelines, embeddings, vector search, pgvector, Qdrant, chatbot-assisted support, trusted retrieval workflows
Databases & Storage
PostgreSQL, MongoDB, Redis, DuckDB, MySQL, SQLite, Qdrant
Cloud & DevOps
AWS, GCP, Docker, Kubernetes, CI/CD, Prometheus metrics, health/readiness endpoints
Security & Data
JWT, RBAC, 2FA (TOTP), CSRF, encryption at rest, blind indexes, ETL pipelines, audit logs, rate limiting, automated testing
Professional Experience
Each role below follows the scope and achievements listed in the CV.
Founder & Software Engineer
2026 - Present
MAAT INNOVATIONS, Novi Pazar
- Design and write a clinic management system with twenty modules that covers the whole provider, not one department: scheduling, clinical records, laboratory (LIS), operating theatre, home visits, stock, equipment, finance and a waiting-room display. See it in use →
- Own every layer: PostgreSQL schema and Alembic migrations, FastAPI services, access control across eight roles, encryption of personal and medical fields at rest with blind-index search, PDF reports, e-mail and SMS.
- Integrations where they matter: two laboratory analysers over ASTM and HL7 v2 (MLLP) through an on-site bridge that keeps results on disk through an internet outage, FHIR R4 export, a fiscal receipt for every payment (ESIR / eFaktura).
- Built the in-house assistant that answers patients from the clinic's own data — local embeddings inside the provider's own database, no external LLM API, and no confident answer to a question the data cannot support.
- Every defect found becomes a regression test that must fail on the old code before the fix lands; 2,400+ checks run in CI on every push, and every run proves it left no test data behind.
Python Engineer
2024 - Present
LINK Group, Belgrade
- Architected and deployed ML-powered applications serving 50,000+ daily active users, and rewrote the data pipelines behind them — a nightly run went from 4 hours to 45 minutes.
- Led development of a predictive analytics platform for higher-quality operational and business decisions.
- Optimized production data pipelines for faster and more reliable analytics delivery.
- Mentored 5 junior developers and introduced best practices for ML model deployment.
- Introduced automated testing in CI, so a broken pipeline is caught before it reaches a client dashboard rather than after.
Python Programming and Machine Learning Instructor
2022 - 2023
Center NIT
- Designed and delivered Python and Machine Learning curriculum for 200+ students.
- Created an automated assessment tool that took routine grading off the instructors.
- Built hands-on projects modelled on real industry work across Python, ML and data workflows.
System Administration
2019 - 2022
City of Novi Pazar
- Managed IT infrastructure for 500+ employees across multiple government departments.
- Implemented automated backups with a restore procedure that was rehearsed, not assumed.
- Took part in digital transformation of citizen-facing services.
- Developed internal Python tools to automate routine administrative workflows.
Private Instructor - Mathematics and Computer Science
2017 - 2019
Belgrade
- Provided personalized instruction to 100+ students in advanced mathematics and programming.
- Developed custom learning materials and interactive coding exercises.
Selected Projects
Project descriptions and technologies aligned to the selected projects section from the CV.
Clinic Management System — MAAT INNOVATIONS
One system for an entire healthcare provider rather than one department: scheduling, clinical records, laboratory, operating theatre, home visits, stock, equipment and finance — my own product, sold as MAAT INNOVATIONS.
- Access control across eight roles, patient data encrypted at rest and searched through blind indexes, an append-only audit trail on every read of a medical record, and integrations over ASTM, HL7, FHIR and fiscal receipts.
- Key technologies: FastAPI, PostgreSQL, pgvector, SQLAlchemy, Alembic, Docker, ReportLab, HL7 v2, ASTM, FHIR R4, JWT, RBAC.
- Business value: one set of numbers across reception, doctors, laboratory and management — nothing entered twice, nothing corrected silently.
Predictive Analytics Platform
Developed an AI-powered predictive analytics platform for historical data analysis, trend forecasting, anomaly detection, and customer behavior analysis.
- Designed ETL workflows with Pandas, NumPy, and Apache Spark, integrated multiple databases, and delivered dashboard reporting with Dash/Plotly.
- Key technologies: FastAPI, TensorFlow, PyTorch, Pandas, NumPy, Apache Spark, PostgreSQL, MongoDB, DuckDB, Dash/Plotly.
- Business value: automates insight generation and improves decision quality through data-backed forecasting.
AI Legal Assistant Platform
Designed an AI legal support platform focused on document understanding, trusted retrieval, and guided assistance workflows.
- Structured around legal document processing, backend services, retrieval pipelines, and user-facing assistant logic.
- Key technologies: AI retrieval workflows, backend services, legal document processing, RAG, vector search, Qdrant, assistant logic.
- Business value: improves information access and helps users navigate document-heavy legal workflows more efficiently.
Drug Interaction Risk Analysis
Created an AI-based system to identify potential drug interactions and predict adverse effects across combined medication use.
- Applied machine learning and NLP techniques for medical text understanding, risk assessment, and safer recommendation workflows.
- Key technologies: Flask, FastAPI, Scikit-learn, TensorFlow, PostgreSQL, spaCy, NLTK.
- Business value: supports pharmaceutical safety, reduces patient risk, and improves medical decision support.
Early Detection of Grapevine Diseases
Developed an AI-powered solution for detecting subtle changes in grapevine leaves and predicting possible plant diseases.
- Implemented deep learning workflows for leaf pattern analysis and used database storage for scalable agricultural monitoring.
- Key technologies: PyTorch, NumPy, Matplotlib, MySQL, Flask.
- Business value: helps vineyards reduce crop loss, improve disease management, and increase yield efficiency.
Datacenter Management API
Developed a REST API for managing datacenter devices and racks with rack-unit capacity validation, power consumption checks, and balanced placement proposals.
- Implemented JWT authentication, RBAC, audit logging, health/readiness endpoints, metrics, rate limiting, optimistic locking, and idempotency support.
- Key technologies: FastAPI, SQLAlchemy, SQLite/PostgreSQL, Docker, JWT, RBAC, Prometheus metrics, audit logs.
- Business value: improves asset control, enforces infrastructure constraints, strengthens traceability, and supports reliable operational decisions.
Education & Certifications
Academic background and certifications presented in the same order as in the CV.
Master's Degree in Computer Science and Mathematics
2022 - 2024State University of Novi Pazar
AI & Python Development
2023 - 2024ITAcademy by LINKGroup, Belgrade
Full-Stack Web Development
2022 - 2023Center NIT, Novi Pazar
State Professional Exam
2021Ministry of Public Administration and Local Self-Government
Bachelor's Degree in Computer Science and Mathematics Education
2012 - 2017State University of Novi Pazar
Computer and Electrical Technician
2008 - 2012Technical School of Novi Pazar
Certifications
- AI & Python Development - ITAcademy by LINKGroup, Belgrade - 2023 - 2024
- Full-Stack Web Development - Center NIT, Novi Pazar - 2022 - 2023