Artificial Intelligence
Diassu provides a wide range of AI/ML data science services. We focus on traditional Linear, Classification, CNN, GAN, VAE, and Bayesian frameworks. These frameworks are used to model uncertainty, update beliefs with new evidence, and make decisions under incomplete information.
Recently, because of all the hype, we have been focusing on LLM/RAG as experts in Watson X, AWS and Azure. So, if you are new to AI/ML, and just need some direction, we offer our AI Transformation Services to figure out how to really transform your business. Let's face it AI/ML and LLM/RAG services are taking the industry by storm.
People are talking about how LLM AI Agents will replace people. We really think that this will take at least another 20-30 years, and our approach portrays this and is a down to earth and practical one where we automate, augment and transform processes. This may include a cloud transformation and/or an AI transformation.
How do we know this?
We use our own software and processes that look nothing like your business already. You will be surprised how innovative we really are. When we are not transforming your business, then we are transforming our own business.
For example, Databricks consulting services includes AI Engineering, MLOps, Data Science, Machine Learning, Data Engineering, EDA (Exploratory Data Analysis), and AI/ML Architecture services. As always, we highly suggest that we start with either John Kruebbe or Robert Clouse Jr. to assess where you are in your AI Journey. Using this valuable service, we then make suggestions on how we can transform your business.
Make sure you are on the leading edge of the AI wave with Diassu Software at your helm. Explore the models and USE CASES we use in our business below...
Our Previous Large Language Model Use Cases
Below you will find the LLM Use Cases that we have previously supported for our customers and internally within our company. We eat our own software because it tastes good especially with Ask Diassu!
Provider |
Model ID |
Context Window |
Parameters (est.) |
Use Case Highlights |
|---|---|---|---|---|
| IBM Granite | ibm/granite-13b-chat-v2 | 8K tokens | 13B | Multi-turn chat, tool use, conversational agents, RAG pipelines, summarization |
ibm/granite-13b-instruct-v2 | 8K tokens | 13B | Structured outputs, classification, summarization, deterministic task execution | |
ibm/granite-3b-8b | 128K tokens | 8B | Long-document summarization, multilingual QA, code generation, hybrid RAG | |
ibm/granite-3b-2b | 128K tokens | 2B | Lightweight multilingual apps, business logic, low-latency inference | |
| IBM Granite | ibm/granite-13b-chat-v2 | 8K tokens | 13B | Multi-turn chat, tool use, conversational agents, RAG pipelines, summarization |
ibm/granite-13b-instruct-v2 | 8K tokens | 13B | Structured outputs, classification, summarization, deterministic task execution | |
granite-3.2-8b-instruct | 128K tokens | 8B | Long-context reasoning, multilingual QA, code generation, RAG, CoT toggle | |
granite-3.2-2b-instruct | 128K tokens | 2B | Lightweight multilingual apps, business logic, low-latency inference | |
granite-vision-3.2-2b | 128K tokens | 2B | Multimodal document understanding, image-to-text, visual RAG | |
granite-guardian-3.2 | 128K tokens | ~3B-A800M (MoE) | Safety-tuned model for compliance, risk detection, and regulated environments | |
| OpenAI | gpt-4o | 128K tokens | ~1.8T (unconfirmed) | Multimodal reasoning (text, vision, audio), coding, chat, advanced RAG |
gpt-3.5-turbo | 16K tokens | ~175B | Fast, cost-effective chat, summarization, classification | |
text-embedding-ada-002 | N/A | ~350M | Embedding generation for search, clustering, vector DBs | |
| Anthropic Claude | claude-3-5-sonnet-20240620 | 200K tokens | Not disclosed | Long-context RAG, forecasting, summarization, image-to-text, code generation |
claude-3-opus-20240229 | 200K tokens | Not disclosed | High-stakes reasoning, financial analysis, research, automation | |
| Amazon Bedrock | amazon.nova-pro-v1:0 | 300K tokens | Not disclosed | Enterprise-grade multimodal workflows (text, image, video), instruction following |
amazon.nova-micro-v1:0 | 128K tokens | Not disclosed | Low-latency chat, summarization, classification | |
amazon.titan-embed-text-v2:0 | N/A | Not disclosed | Text embeddings for semantic search, RAG, and clustering |