Muhammad Salman Siddiq

Academic

Career

Verbatim text

Muhammad Salman Siddiq
Cell: 03319994652 |  Email: salmansiddiq566@gmail.com
LinkedIn: https://www.linkedin.com/in/muhammad-salman-siddiq-43271b250
Address: HOUSE NO.313 , Street 17, phase 7 bahria town, rawalpindi , Rawalpindi , Pakistan
PROFESSIONAL PROFILE
I'm a Software engineer with hands-on experience building production systems that scale. Over the past year, I've designed and
deployed async, high-performance APIs using FastAPI and PostgreSQL, handled OAuth integrations across multiple providers, and
built AI-powered automation platforms from architecture to deployment. I work comfortably with Docker, microservices, and CI/
CD pipelines, and I've deployed containerized applications on GCP with automated workflows.
My work spans both SQL (PostgreSQL with SQLAlchemy async patterns) and NoSQL databases (MongoDB), and I've optimized
systems for performance through query tuning, async processing, and background job orchestration. I thrive in fast-paced
environments—writing clean, testable code, debugging complex issues, and shipping reliable software quickly. I'm proficient in
Python, comfortable with Linux and bash scripting, and experienced with REST APIs, webhooks, and microservices architecture.
I bring strong problem-solving skills, a collaborative mindset, and a track record of owning projects end-to-end. Whether it's building
LLM-powered workflows, integrating third-party APIs, or troubleshooting production issues, I focus on delivering systems that work
well and solve real problems.
EDUCATION
Bachelors of Engineering in Software Engineering
SEECS , RAWALPINDI , 3.54 (2026)
INTERNSHIP EXPERIENCE
Antematter
01-Jun-2025 - 01-Oct-2025
• Contributed to building a data analysis co-pilot for hospitality industry by developing and testing parts of an LLM- driven multi-agent
system, achieving sub-60s response times for SQL + Python tasks. • Automated testing and evaluation of agent performance using
custom scripts, running agentic evaluations to spot failure cases and improve reliability in real-world SQL + Python workloads. • Built
a full-stack subtitle scraping tool (FastAPI, Next.js) to extract and process YouTube captions; used threads and worker pools to
significantly reduce batch processing time. • Set up n8n automation workflows linking GitHub, Jira, and Discord; streamlined team
updates by auto-syncing code pushes with ticket status changes and notifications.
FINAL YEAR PROJECT
Adaptive Web Interaction: Leveraging RL for Comprehensive Action Support
We are building a multi-agent AI system that can navigate websites based on natural language instructions from users. For privacy
and computational efficiency, we are fine-tuning open-source small language models (SLMs) using supervised learning and
advanced reinforcement learning techniques. As part of this effort, we are also developing automated pipelines to generate a large-
scale training dataset using LLMs.
TECHNICAL EXPERTISE

AI enrichment