Looking for a job

Syed Muhammad Taha

@syedmuhammadtaha

Message

AI/ML Engineer at Brainbox Automations deploying safety-guardrailed AI systems on GCP for 300+ athletes.

Pakistan

Message

At Brainbox Automations, I deployed a safety-guardrailed AI coaching backend on GCP Cloud Run for more than 300 minor and adult athletes. I built content-moderation guardrails for offensive and sexual content, and optimized the GitHub Actions deployment pipeline from about two minutes to 10 seconds.

I also engineered SQL-aware agents for a Supabase-hosted RAG chatbot across more than 50 tables, improving query accuracy and cutting response time by 50%. In partnership with Earlibird, I built customer-facing chat and VAPI voice AI for brand queries, appointment booking, and product search.

At Epistemy, I led end-to-end development of an AI tutoring platform, including its Nest.js backend and Next.js frontend. I also architected a Redis/BullMQ task queue for multi-agent workflows and maintained unit test coverage above 90%.

My research spans foundation-model fine-tuning for wireless sensing and clinical factual verification. At MachVis, I proposed an LLM-based framework for checking pathology-report accuracy and curated a challenge dataset of more than 100 gigapixel whole-slide images.

Experience

Work history, roles, and key accomplishments

Current

Research Intern

Current

IPT Lab, NUST

Jun 2026 - Present (3 months)

Fine-tuning pre-trained wireless/RF foundation models for downstream sensing tasks, including human activity recognition and beam prediction. Applying parameter-efficient fine-tuning techniques to adapt models under limited compute and labeled data.

Fine Tuning PEFT Foundation Models HAR Logs Beam Prediction

Current

AI Engineer

Current

Brainbox Automations

Dec 2025 - Present (9 months)

Deployed a safety-guardrailed AI coaching backend on GCP Cloud Run serving 300+ athletes, and engineered SQL-aware agents for a Supabase-hosted RAG chatbot. Partnered with Earlibird to build a customer-facing chat and VAPI voice AI system handling 200+ daily interactions.

GCP Cloud Run Docker fastAPI Vertex AI GitHub Actions CI CD RAG Supabase SQL VAPI AWS Lambda Gemini API MapReduce

Undergraduate Research Intern

MachVis

Oct 2025 - Dec 2025 (2 months)

Proposed a novel LLM-based factual verification framework to evaluate the clinical accuracy of pathology reports. Curated a challenge dataset of 100+ gigapixel whole-slide images to assess model robustness.

LLM Fact Checking Pathology Whole Slide Imaging Systems Dataset Curation

AI Fullstack Engineer

Epistemy

Sep 2025 - Dec 2025 (3 months)

Led end-to-end development of an AI tutoring platform, architecting a Nest.js backend and Next.js frontend. Architected an event-driven task queue with Redis/BullMQ and enforced SWE standards with CI pipelines and pre-commit hooks.

NestJS Next.js Supabase Redis BullMQ CI CD Unit Testing

Machine Learning Intern

CogniMind AI

Feb 2025 - Apr 2025 (2 months)

Automated MLOps by deploying dockerized Apache Airflow and designing 5+ DAGs, reducing manual intervention by 60%. Improved model performance by enhancing VLM extraction accuracy and optimizing retrieval/inference pipelines.

Airflow Docker GitHub Actions CI CD Prompt Engineering Quantization HNSW

Machine Learning Intern

RapidsAI

Sep 2024 - Dec 2024 (3 months)

Developed and integrated a RAG-based chatbot with contextual session management, using Chain-of-Thought prompting to reduce errors by 50% and a multi-model query router to cut OpenAI API costs by 50%.

RAG Streamlit fastAPI Chain Of Thought OpenAI API Query Routing

Education

Degrees, certifications, and relevant coursework

National University of Sciences & Technology

Bachelor of Science, Computer Science

2022 - 2026

Grade: 3.5/4.0

Activities and societies: Executive @ NUST Science Society, Team Lead @ GDSC

Pursued a Bachelor of Science in Computer Science, achieving a cGPA of 3.5/4.0. Ranked 20th in a class of 127, served as Executive at NUST Science Society and Team Lead at GDSC.