Can Sahin — Software Engineer | Backend and Full-Stack

Work Experience

IT Analyst (Co-op) - University of Ottawa (University of Ottawa IT Initiatives Support)

Duration: Sept 2023 to Apr 2024
Summary:

• Supported IT architecture governance and improved operational efficiency at the University of Ottawa by automating processes and coordinating governance activities
• The role involved enhancing meeting distribution accuracy and exploring generative AI adoption opportunities

Responsibilities:

• Developed a Python script to eliminate manual corrections in meeting distribution lists by cross-referencing group memberships and flagging discrepancies.
• Reviewed business and technical architecture documents to support architecture governance.
• Coordinated weekly Architecture Working Group (AWG) and monthly Architecture Review Board (ARB) sessions with project managers.
• Researched and presented generative AI adoption opportunities to IT architects and stakeholders.

Technologies: Python

Software Developer (Co-op) - March Networks (Compliance Automation Tools)

Duration: Jan 2023 to Apr 2023
Summary: Developed automation tools to streamline compliance review processes for legal and R&D teams at March Networks, significantly reducing manual effort and improving efficiency in license and component review workflows.
Responsibilities:

• Automated Black Duck report generation and formatting, replacing manual downloads and Excel macros.
• Built a Python tool to compare component, version, and license IDs across release versions to identify medium and high risk licenses.
• Optimized matching logic to reduce comparison runtime by approximately 50% on large releases.

Technologies: Python

Software Engineer (Go API Gateway)

Summary:

• Developed a high-performance Go API gateway capable of sustaining 13
• 7k requests per second with features including Redis-backed atomic rate limiting, circuit breakers for upstream services, and per-route metrics monitoring
• The gateway was deployed live on Render and monitored with Prometheus and Grafana

Responsibilities:

• Implemented Redis-backed rate limiting enforced atomically via Lua scripts.
• Developed circuit breakers to handle failing upstream services gracefully.
• Instrumented per-route metrics and integrated monitoring with Prometheus and Grafana.
• Deployed and maintained the gateway on Render platform.

Technologies: Go, Redis, Lua, Prometheus, Grafana, Render

Software Engineer (Job Search RAG Pipeline)

Summary:

• Built a Retrieval-Augmented Generation (RAG) system over 32 live Canadian job postings to improve job search accuracy
• The system combined BM25 and dense vector search with Reciprocal Rank Fusion and cross-encoder reranking, achieving high citation accuracy and composite score

Responsibilities:

• Implemented combined BM25 and dense vector search techniques with Reciprocal Rank Fusion.
• Developed cross-encoder reranking to reduce candidates from 20 to 5.
• Integrated an LLM-as-judge to verify claims and flag hallucinations.

Technologies: Python, LLM, BM25, Dense Vector Search

Full-Stack Developer (Gym Progression Tracker)

Summary:

• Delivered a full-stack application to automate progressive overload decisions for gym users
• The app featured a domain engine that auto-increments load on success and applies deload after failures, with JWT authentication and error tracking, deployed on Render and Vercel

Responsibilities:

• Developed frontend using React and TypeScript.
• Built asynchronous backend APIs with FastAPI and PostgreSQL database.
• Implemented domain logic for progressive overload and deloading.
• Deployed application on Render and Vercel with JWT authentication and Sentry error tracking.

Technologies: React, TypeScript, FastAPI, PostgreSQL, JWT, Sentry, Render, Vercel

Backend Developer (Semantic Caching Layer for LLM APIs)

Summary:

• Created a FastAPI middleware proxy to reduce redundant LLM API calls and latency by serving cached responses for semantically similar requests
• The system used sentence-transformers for embedding, Qdrant for vector storage, and was instrumented for monitoring

Responsibilities:

• Built FastAPI middleware proxy for caching LLM API responses.
• Embedded queries using sentence-transformers.
• Stored vectors in Qdrant with cosine similarity thresholds and TTL expiration.
• Dockerized the service and integrated Prometheus and Grafana for monitoring.

Technologies: FastAPI, Python, sentence-transformers, Qdrant, Docker, Prometheus, Grafana

Backend Developer (Spring Boot RabbitMQ Order Service)

Summary:

• Developed a Java 21 order microservice using Spring Boot and RabbitMQ to ensure duplicate-free message processing with idempotent consumers, dead-letter queue, and exponential-backoff retry mechanisms
• The service was validated with unit and integration tests

Responsibilities:

• Implemented idempotent message consumers backed by a processed-events table.
• Configured dead-letter queue and exponential-backoff retry strategies.
• Conducted unit and integration testing against Dockerized PostgreSQL and RabbitMQ.

Technologies: Java 21, Spring Boot, RabbitMQ, PostgreSQL, Docker

Education

• Honours BSc Computer Science, Co-op Option, University of Ottawa
  Graduated Cum Laude. Dean's Honour List (Winter 2025).
  Sept 2020 to Apr 2025
  

Certification

• AWS Cloud Solutions Architect (Coursera)