深度学习 Jobs

I need an end-to-end AI solution that lets a runway-patrol drone spot even the tiniest foreign object debris (FOD) in real time. The airframe already carries LiDAR, an infrared sensor and an electro-optical (EO) camera; what is missing is the software that fuses those feeds, classifies debris and reports exact GPS position so ground crews can clear it fast. Scope of detection • Metallic pieces, plastic fragments and loose stones—down to roughly 1 mm in size. • Operation must be reliable in both daylight and nighttime conditions. • Accuracy target is set to a high threshold; false positives have to stay low while recall stays near perfect. Key tasks 1. Build or adapt a multi-sensor fusion pipeline that consumes LiDAR point clouds, IR imagery and EO frames. 2...

Please read the following instructions carefully before starting. 1. Sample Images The client has provided 10 sample images. Since all of the samples follow the same annotation standard, I have divided them among the team to save time. Your assigned sample image should be treated as the reference while annotating your assigned batch. 2. Requirements - Draw Polygon annotations for all required objects visible in the image. - Apply Key Segmentation only for Broad Leaf and Weed, exactly as demonstrated in the sample images. - Do not use any different annotation style. Follow the sample as closely as possible. 3. Annotation Quality - Polygon boundaries must be tight and accurate. - Do not include unnecessary background. - Cover the complete visible area of each object. - If multiple obj...

I have mapped out a consumer-facing AI platform that blends natural conversation, real-time insights, and intelligent automation into a single web & mobile experience. The roadmap is clear, the market niche is validated, and I’m now ready to turn the concept into an MVP and beyond alongside a technical cofounder who brings deep expertise in AI. Core vision The product centres on three pillars: seamless user interaction and support through robust NLP, data analysis with intuitive visualisation, and automated decision-making that genuinely helps everyday users, not just enterprises. Under the hood we will combine Natural Language Processing, Computer Vision where it adds tangible value (e.g. image-based recommendations), and classic plus deep Machine Learning models to personalise...

I am building a fully self-hosted, real-time voice assistant on top of LiveKit. The system must ingest the caller’s audio stream, perform on-premise voice recognition, run natural-language processing with contextual comprehension, and respond intelligently as a first-line customer-support agent—all without relying on any third-party SaaS such as Vapi, ElevenLabs, or similar hosted APIs. Open-source or locally runnable models are fine; cloud subscriptions are not. Core flow • LiveKit delivers the audio track in real time. • Your pipeline converts speech to text, keeps multi-turn context, and routes it through an NLP layer capable of understanding intent and maintaining conversation state. • The agent then generates the textual reply that my existing interfa...

I need help configuring a Hikrobot camera using python whenever it flags a vehicle incorrectly, drives an external alarm output. In practice that means: • Monitoring the confidence/result coming from the camera (SDK, ONVIF events or RTSP overlay—whatever is most reliable) and deciding when a detection should be considered “wrong”. • As soon as a wrong detection is confirmed, toggling the camera’s I/O port (or a small companion relay module if the onboard I/O proves too limited) to fire the alarm signal. A concise set-up guide and any scripts, firmware tweaks or parameter files you create must be supplied so I can replicate the installation on additional units. Please include short test footage or logs proving that the alarm is triggered only on false...

The project centers on building a production-ready TensorFlow 2.x model that classifies tabular data delivered to us through an internal API. I have the API specifications and sample payloads ready; you will turn those streams into a clean training pipeline, engineer the right features, and iterate until the classifier meets our performance targets in real-world tests. Scope of work • Data pipeline – pull the API data, handle preprocessing, and produce TensorFlow-friendly datasets for train/val/test splits. • Model development – design, train, and tune a deep learning architecture suitable for tabular inputs (e.g., wide & deep, Transformer, or other proven structures). • Optimization – experiment with hyperparameters, regularization, and callback...