Syed Mirak Wajahat Kirmani

AI-Powered Document Analysis and Question Retrieval System
Tools: Python, LangChain, Streamlit, FAISS, OpenAI

• Developed a retrieval-augmented generation (RAG) system to analyze documents and extract question-based insights, improving document comprehension efficiency by 20%.
• Leveraged LangChain to modularize pipeline steps, including chunking, embedding, and query routing.
• Integrated with Streamlit for interactive visualization and transparency.

Interactive RAG with Human-in-the-Loop Feedback System And Few-Shot Generalization
Tools: Python, MongoDB, LangChain, FAISS

• Designed a feedback-enabled RAG framework that improved answer relevance based on user rankings.
• Engineered semantic similarity-based few-shot generalization to boost answer quality for unseen queries.

Human Movement Detection System
Tools: Python, OpenCV, TensorFlow, Streamlit

• Developed a real-time human movement detection system using computer vision techniques and deep learning models.
• Utilized OpenCV for video stream processing and pre-trained CNN models for accurate motion recognition in surveillance footage.
• Designed a lightweight Streamlit interface to visualize alerts and detection results live, enabling non-technical users to monitor movements interactively.
• Achieved over 92% detection accuracy in controlled environments and facilitated model deployment in low-latency settings.

SAS to PySpark Migration for Sequence Prediction
Tools: SAS, PySpark, Bi-directional LSTM, Databricks

• Implemented a Bi-Directional LSTM model for sequential event prediction, improving accuracy by 18% over traditional time-series models.
• Automated the data ingestion and transformation workflow using PySpark and Databricks jobs, reducing preprocessing time by 45%.

Contact Center – FCR & NPS Optimization
Tools: Python, Excel, SLM

• Analyzed multi-LOB feedback to identify root causes of dissatisfaction.
• Delivered GenAI insights and recommendations that improved customer KPIs by 15%.

Time Series Modularization for Forecasting
Tools: Python (ARIMA, LSTM, Prophet)

• Built a reusable notebook for automating data cleaning, seasonality detection, and forecasting.
• Reduced manual effort and improved forecast consistency across business units.

Business Process Optimization using SQL, PySpark, and ML
Tools: SQL, PySpark, XGBoost, Databricks

• Built an end-to-end ML pipeline for anomaly detection in retail processes using PySpark and SQL.
• Achieved 91% precision using XGBoost, improving decision-making in business operations.

Reporting Automation using Power BI
Tools: Power BI, Excel

• Migrated 35 Excel-based reports to Power BI, reducing reporting time by 82%.
• Enhanced real-time CX monitoring with dynamic dashboards.