A curated portfolio of autonomous agent networks, enterprise RAG engines, deep learning models, and MLOps deployments engineered by Rohit.
State-aware multi-agent system built using CrewAI and LangGraph. Automatically monitors inventory deficits, queries Qdrant vector memory, and issues PO re-routing calls.
Hybrid search engine combining Qdrant dense vector embeddings, BM25 sparse keyword index, and Cohere Reranking layer over 4.5M medical papers.
PyTorch deep autoencoder streaming transaction analysis capable of identifying financial fraud patterns in under 140 milliseconds.
Private GPU infrastructure setup running open-weight 70B LLMs with vLLM PagedAttention, achieving 10x higher token throughput at zero data leak risk.
Enterprise document extraction pipeline processing 100,000+ legal contracts with Unstructured parser, Pinecone vector store, and Claude 3.5 Sonnet.
Multi-agent software engineering assistant that reads pull requests, executes static code linters, runs unit test suites, and posts audited code reviews.
Autonomous scraping crew powered by Playwright and CrewAI that monitors competitor pricing, product updates, and sentiment changes 24/7.
BioBERT-based clinical triage engine prioritizing emergency patient symptoms and cross-referencing electronic health records securely.
Autonomous financial risk management agent evaluating Value-at-Risk (VaR) thresholds and auto-executing portfolio hedges under volatility spikes.
Multi-agent support system reading complex technical tickets, querying internal knowledge bases, and executing refund/tier escalations autonomously.
Time-series forecasting model combining XGBoost and PyTorch LSTM to predict customer re-order intervals with 94.2% precision.
Generative Adversarial Network (GAN) pipeline producing anonymized synthetic tabular data for HIPAA and GDPR compliant ML training.
Let’s discuss your data architecture, vector databases, and agent tool execution requirements directly with Rohit.
Discuss Your Project