Curriculum Vitae

Algorithm Engineer · CV & NLP

Building intelligent vision & language systems with production-ready quality, low latency, and measurable impact.

Open to Opportunities Deep Learning MLOps · Distributed · Research-to-Prod

About

Experienced algorithm engineer specialized in computer vision and natural language processing. Skilled at designing end-to-end ML pipelines—from data curation, modeling, and evaluation to deployment on GPU/CPU clusters. Proven track record in latency-sensitive inference, model compression/distillation, and delivering user-facing AI features.

Education

Master of Software Engineering
2020 - 2023 · XX University

Focus: Deep Learning, Distributed Systems, MLOps
Bachelor of Computer Science
2016 - 2020 · XX University

Projects on image recognition, information retrieval

Work Experience

Senior Algorithm Engineer · TechCorp
2022 - Present

Led CV & NLP model development for large-scale content platforms; delivered real-time moderation and recommendation signals under strict SLA; improved inference cost by 30% through pruning + distillation.
Algorithm Engineer · Startup AI Lab
2020 - 2022

Shipped multimodal search features combining vision-language embeddings; built data pipelines, evaluation harnesses, and A/B tests to validate ranking quality.

Projects

Multimodal Retrieval System

2023 · Vision-Language · Milvus/FAISS

Designed contrastive training for image-text embeddings; implemented ANN indexing with shard routing; reduced query latency to <80ms P95 at 10M+ items.

Real-time Content Safety

2022 · CV/NLP · Distillation/Pruning

Built cascaded models (OCR + NSFW + toxic text); applied quantization-aware training, cutting GPU usage by 25% while keeping F1 parity.

Edge-friendly Vision Inference

2021 · On-device · TFLite/NNAPI

Converted and optimized detection/classification models for mobile; introduced dynamic input batching and pre/post-processing fusion to achieve 2x throughput.

Publications & Patents

Low-latency Vision Pipelines

2024 · Conference (under review)

Proposed a modular pipeline with adaptive resolution and operator-level caching; demonstrated 18% latency reduction on large-scale benchmarks.

Efficient Transformer Distillation

2023 · Journal Submission

Introduced dual-stage distillation for vision-language transformers, preserving alignment quality with 40% parameter reduction.

Patent: Content Understanding Framework

Filed 2023

System and method for unified multimodal tagging and retrieval with adaptive model routing across heterogeneous hardware.

Contact

Email: yourname@example.com
Phone: +86 138-0000-0000
LinkedIn: linkedin.com/in/yourname
GitHub: github.com/yourname