genai-pinnacle-plus

5 prompt engineering techniques on ABN AMRO financials via Claude Vision API (zero-shot, CoT, role-based, JSON, multi-image)

AI Agent industry report (.1B→7B market) auto-generated as .docx using Python stdlib + multi-LLM RAG | AV GenAI Pinnacle Plus

4 AutoGen patterns: Reflection loop · FunctionTool BMI · StateFlow portfolio manager · Multi-modal bill analyzer

LangGraph + Claude Sonnet + MemorySaver: persistent competitive intelligence agent for retail stores

LangChain AgentExecutor: live weather (WeatherAPI) + DuckDuckGo attractions → structured travel summary

CrewAI sequential: Logistics Analyst → Optimization Strategist — VRP routing + ABC/XYZ inventory roadmap

Dual-task DNN on CPCB water quality data: WQI regression + classification | TensorFlow/Keras | AV GenAI Pinnacle Plus

Advanced RAG: LIC policy (3 embeddings x 2 LLMs) + Coca-Cola 10K 2014–2023 with Cohere reranker + SubQuestion engine

Flask REST API: live stock data, OHLCV history, Sharpe/drawdown/MA analytics + BUY/HOLD/SELL | AV GenAI Pinnacle Plus

Advanced CrewAI: learning path generator (SerperDev + custom tool) + code debugger (CodeInterpreterTool)

BPE tokenizer from scratch on WikiText-2: 30K vocab, special tokens, TemplateProcessing → PreTrainedTokenizerFast

Health classification ML pipeline: 4 models, feature engineering, clinical data cleaning | AV GenAI Pinnacle Plus

LangGraph StateGraph: math tool-use loop + multi-agent router (web/RAG/LLM) with typed state | AV GenAI Pinnacle Plus

PDF → LLM summary → MCQ quiz generator via LangChain SequentialChain | AV GenAI Pinnacle Plus

1.6M tweet sentiment: TF-IDF baseline → PyTorch RNN/LSTM/GRU/BERT comparison | AV GenAI Pinnacle Plus

Structured business analysis: customer retention problem decomposition using GenAI-assisted methodology

Full RAG: FAISS + MiniLM embeddings + MMR retrieval + Flan-T5-Large — runs 100% local, no API key needed

QLoRA fine-tuning: 4-bit NF4 quantization + LoRA adapters on BERT — trains only ~0.1% of 110M params

4 ML assignments: healthcare x3 + NYC Taxi 729K row regression | scikit-learn, XGBoost | AV GenAI Pinnacle Plus

Custom ReAct agent from scratch (no framework): Groq LLM plans, Tavily searches, LLM synthesizes report