Machine Learning Engineer Internship

Decrypt’s Art, Fashion, and Entertainment Hub. Discover SCENEIf you are a regular social media user, then you may have stumbled upon videos like this:China AI-Generated videos depicting MAGA supporters - including Trump himself - working in warehouses sewing and manufacturing are going viral after the trade war between Beijing and Washington kicked off.pic.twitter.com/XP1jd4oyLC — The American Way With Sheila Kay (@usasheilakay) April 11, 2025Or this:Trump taking a 30 min lunch break from the factory work at the fat factory model he created that will make America great again pic.twitter.com/941SMKojGY — Utamadush 🦇 🔊 ➕ (@utamadush) April 9, 2025Or thisHilariousChina is making some real stuff though AI right now.China vs USA was heating up. MAGA#TrumpTariffs #stockmarketcrash #tariff pic.twitter.com/XYrT7tSIfZ — Anshul Garg (@AnshulGarg1986) April 11, 2025Thanks to the U.S. trade war, viral videos depicting overweight Americans toiling in fictional sweatshops have exploded across TikTok and X, racking up millions of views. These clips show exhausted workers stitching clothes in dismal factory conditions—a satirical jab at Trump's promise to bring manufacturing jobs back to America through hefty tariffs on Chinese goods.One 32-second video created by TikTok user Ben Lau has been viewed millions of times, showing AI-generated Americans in dire factory conditions, complete with traditional Chinese music and ending with a snarky "Make America Great Again" slogan.AI has become the new political cartoon, a powerful tool for political protest, and savvy users are coming up with creative ways to use their favorite models as means to spread a message.Ever wondered how they do it? It’s actually pretty easy. All you need is a powerful enough PC—or be willing to spend a few bucks/euros/pesos/pounds/yuan—to bring your ideas to life

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More. The race to expand large language models (LLMs) beyond the million-token threshold has ignited a fierce debate in the AI community. Models like MiniMax-Text-01 boast 4-million-token capacity, and Gemini 1.5 Pro can process up to 2 million tokens simultaneously. They now promise game-changing applications and can analyze entire codebases, legal contracts or research papers in a single inference call.At the core of this discussion is context length — the amount of text an AI model can process and also remember at once. A longer context window allows a machine learning (ML) model to handle much more information in a single request and reduces the need for chunking documents into sub-documents or splitting conversations

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More. Researchers at Together AI and Agentica have released DeepCoder-14B, a new coding model that delivers impressive performance comparable to leading proprietary models like OpenAI’s o3-mini. Built on top of DeepSeek-R1, this model gives more flexibility to integrate high-performance code generation and reasoning capabilities into real-world applications. Importantly, the teams have fully open-sourced the model, its training data, code, logs and system optimizations, which can help researchers improve their work and accelerate progress.Competitive coding capabilities in a smaller packageThe research team’s experiments show that DeepCoder-14B performs strongly across several challenging coding benchmarks, including LiveCodeBench (LCB), Codeforces and HumanEval+.“Our model demonstrates strong performance across all coding benchmarks… comparable to the performance of o3-mini (low) and o1,” the researchers write in a blog post that describes the model.Interestingly, despite being trained primarily on coding tasks, the model shows improved mathematical reasoning, scoring 73.8% on the AIME 2024 benchmark, a 4.1% improvement over its base model (DeepSeek-R1-Distill-Qwen-14B). This suggests that the reasoning skills developed through RL on code can be generalized effectively to other domains.Credit: Together AIThe most striking aspect is achieving this level of performance with only 14 billion parameters. This makes DeepCoder significantly smaller and potentially more efficient to run than many frontier models.Innovations driving DeepCoder’s performanceWhile developing the model, the researchers solved some of the key challenges in training coding models using reinforcement learning (RL).The first challenge was curating the training data