Motivated AI Engineer with strong foundation in machine learning, data preprocessing, and model evaluation. Experienced in Python, scikit-learn, and deep learning architectures. Quick to adopt new tools and frameworks, w…
Motivated AI Engineer with strong foundation in machine learning, data preprocessing, and model evaluation. Experienced in Python, scikit-learn, and deep learning architectures. Quick to adopt new tools and frameworks, with a focus on writing clean, reproducible code. Eager to contribute to cross-functional teams and solve challenging problems using cutting-edge AI techniques. Committed to delivering scalable AI solutions that align with business objectives and user needs. Passionate about staying current with latest AI research and translating theory into practical applications.
<p>1.Data preprocessing: Clean, organize, and prepare datasets for training machine learning models.</p><p>2.Model development: Build and experiment with machine learning and deep learning algorithms.</p><p>3.Model training: Adjust hyperparameters and optimize models for better performance.</p><p>4.Model evaluation: Test models using metrics such as accuracy, precision, recall, and F1-score.</p><p>5.Deployment support: Assist in deploying models to production environments or cloud platforms (AWS, Azure, GCP).</p><p>6.Documentation: Write technical reports and maintain clear documentation of processes and results.</p><p>7.Collaboration: Work closely with data scientists, software engineers, and product teams to deliver AI solutions.</p><p>8.Research: Explore new tools, frameworks, and research papers to apply innovative approaches.</p><p>9.Experiment tracking: Use tools like TensorBoard or MLflow to track experiments and analyze outcomes.</p>
<ul><li>Fine-tuned XLM-R with SetFit to classifyNepali legal documentsusing only 50 labeled examplesper class, achieving 83% F1 score</li><li>Reduced labelingeffort by 80% via active learning, outperforming zero-shot GPT-3.5 (67% F1) and mBERT (71% F1)</li><li>Deployed a Gradio UI for real-timedocument classification, enablingnon-technical users to upload and classify instantly</li></ul>
<ul><li>Built an LSTM autoencoder in PyTorch achieving94% precision and 91% recall on Numenta benchmark, outperforming isolation forest by 18%</li><li>Deployed via FastAPI with Kafka streaming, processing 2000 events/second at
<ul><li>Developed a LangGraph-based agentthat decomposes complexqueries into 3–5 sub-tasks with automated web search and summarization, delivering cited reports in under 2 minutes</li><li>Integrated TavilySearch API and FAISS vectormemory to reduceredundant searches, cuttingAPI costs by 35% while maintaining 92% citation accuracy</li><li>Implemented robust error handlingwith exponential backofffor rate limits and broken links, ensuring reliable real-world operation</li></ul>
<ul><li>Built a collaborative multi-agent system (AutoGen) with 4 specialized agents for data fetching, sentiment analysis, risk assessment, and trade execution, reducing decision latencyby 40%</li><li>Integrated Yahoo Finance and News APIs for real-timeprice and sentimentanalysis, achieving 8% simulated portfolio return over 3 months</li><li>Visualized agent communication and trade decisions via Streamlit dashboard, outperforming S&P 500 by 3% in backtestin</li></ul>
The AI-first Recruitment Operating System that unifies and automates the entire hiring lifecycle for modern enterprises worldwide.
Dr Job FZ LLC. 2026 © All Rights Reserved
