Open to opportunities

Kulsoom Naqvi

@kulsoomnaqvi

Message

AI/ML engineer building low-latency, multimodal edge and robotics AI systems.

United States

Message

What I'm looking for

I seek roles developing production-grade, low-latency multimodal AI and robotics systems with strong engineering practices, containerized microservices, and opportunities to impact real-time decision-making at the edge.

I am an AI/ML engineer with 5+ years designing and deploying multimodal systems across speech, vision, and language, focused on low-latency edge and robotics applications. I specialize in building scalable, efficient pipelines using C++, Python, containerized microservices, and RAG-driven retrieval with human-in-the-loop oversight.

I have led production projects including an AI receptionist with GPT-4, compliance verification systems, multimodal voice agents using Whisper.cpp/Llama.cpp/Piper, and optimized Databricks pipelines for 10M+ records. I deliver measurable impact—reduced verification time by 60%, cut clinician search time by 40%, and achieved 92% extraction accuracy on large-scale prescription NLP.

Experience

Work history, roles, and key accomplishments

Current

Senior Software Engineer

Current

QTEX

Jan 2023 - Present (3 years)

Built AI receptionist and compliance verification systems using GPT-4, LangChain, and OCR that automated client intake and reduced manual verification time by 60%; optimized Databricks pipelines to preprocess 10M+ records, cutting data prep by 65%.

GPT 4 Lang Chain Databricks RAG Faiss SpaCy Med7 Docker gRPC

Senior Machine Learning Engineer

Medtran

Jan 2020 - Jan 2023 (3 years)

Designed and deployed multimodal AI agents using Whisper.cpp, Llama.cpp and Piper/Kokoro for real-time robotics interactions and built custom RAG pipelines with containerized orchestration, reducing edge inference latency by 40%.

Kokoro RAG Pipecat Ollama Docker gRPC

Software Engineer

Procore

Jan 2019 - Jan 2020 (1 year)

Supported development of noise-robust STT and conversational AI models, automated preprocessing and model serving workflows to cut inference time by 25%, and collaborated on Docker-based reproducible deployments.

Speech Recognition Docker data preprocessing Git NLP Deep Learning PyTorch TensorFlow

Education

Degrees, certifications, and relevant coursework

Coursera

Professional Certificate, Machine Learning

2020 - 2020

Completed the Machine Learning Specialization consisting of coursework on core ML algorithms, model evaluation, and practical application of ML techniques.

University of Virginia

Bachelor of Science, Computer Science

2015 - 2019

Completed a Bachelor of Science in Computer Science with coursework and projects focused on machine learning, speech and language processing, and systems for deploying ML applications.