Ankita Singh

As a Generative AI Engineer, I led the development of two real-world LLM-powered systems that combine Natural Language Processing (NLP), LangGraph workflows, and fine-tuned foundation models to solve practical problems in education and healthcare domains. 🔹 SmartMail Agent (Email Intelligence System) Built an AI-driven email processor that automatically categorizes 1,000+ school-related emails into predefined categories like Academics, Transport, and Fees using Transformers, Sentence Embeddings, and ChromaDB. Integrated Pinecone vector search and APScheduler to manage follow-up reminders based on semantic intent and unresolved threads. Delivered 92%+ classification accuracy and streamlined school communication workflows using FastAPI, MongoDB, and JWT-secured APIs. 🔹 Medical Health Chatbot (Symptom-to-Treatment LLM System) Designed a multi-phase chatbot using LLaMA 3.1 and 3.2, capable of extracting user symptoms, predicting diseases, recommending medicines, and generating personalized treatment plans. Fine-tuned LLaMA 3.2 1B on curated 1mg.com product data for disease-to-medicine mapping, boosting recommendation precision by 85%. Implemented LangGraph orchestration and memory checkpointing for stateful, multi-turn medical conversations with structured outputs in JSON and natural language. 🔬 Model Evaluation Pipeline (Ollama) Developed an end-to-end evaluation pipeline using pandas, scikit-learn, ollama, and matplotlib to benchmark disease prediction models. Automated data preprocessing, inference, and visualization with a modular folder structure for reproducible ML experimentation.

Akira AI: Authored complete test case documentation and bug reports to ensure functional accuracy. 🔹 Xenonstack Product Testing: Validated redirect links and maintained accuracy on coverage tracking sheets. 🔹 TYP: Conducted E2E testing for WhatsApp SMS delivery scenarios; tested and documented issues with transport order flows and warehouse handling. 🔹 Xenonstack Website: Performed E2E and UI testing on multiple sections; automated Lighthouse performance insights, validated URL redirects, and contributed to SEO testing. 🔹 ALP (Automated Landing Platform): Created test and vulnerability documentation, wrote manual test cases based on Figma, and ensured end-to-end functional security. 🔹 Xenonstack AI: Conducted multi-level testing on staging and deployment; automated testing with Cypress; validated UI content and ensured bug resolution. 🔹 Stacklabs (Playwright Automation): Developed and maintained automated E2E scripts for login, dashboard, member, and deployment flows; created feature-wise bug reports, test documentation, and Loom demos. 🔹 Experimental Saturdays: Integrated Okta authentication with React; documented architecture and implemented role-based access. Built and deployed a GitHub Action Bot for PR automation. 🔹 QA Agent Project: Built a full-stack testing agent using Playwright, FastAPI, PostgreSQL, and LangGraph. Integrated AI-powered flakiness detection with Groq LLM, real-time test reporting with WebSockets, and automated debugging with trace viewer and video logs. Tools & Technologies Languages: Python, JavaScript (React.js), SQL, CSS Frameworks: Playwright, Cypress, FastAPI, LangGraph Testing: Manual Testing, E2E Testing, Bug Tracking, Vulnerability Testing Tools: Lighthouse, Postman, SQLAlchemy, Alembic, GitHub Actions, Okta, WebSockets Platforms: Xenonstack AI, ALP, Stacklabs, TYP, Akira AI, Okta, GitHub Reporting: HTML Test Reports, Structured Bug Sheets, Performance Insights

Worked on diverse real-world projects during my internship at XenonStack, gaining hands-on experience in cloud infrastructure, backend development, RPA, QA, and data engineering. 🔹 Amdocs – Cloud & Data Management Updated AWS instance sheets using Cloud Grazer and scripts for Defender versions, IPs, and clusters. Created structured spreadsheets to handle lookup and subnet use cases. Synced Wireless 5G data with JSON files and verified missing IPs. Wrote a Boto3 script to fetch user data from Cloud9 and EKS for user management. 🔹 Akira AI – Backend Development Built APIs for user and customer onboarding. Managed customer data efficiently and documented APIs using Swagger. Collaborated with frontend team for API integration and testing. 🔹 Gen AI Platform – Backend Support Developed login/signup functionality and dashboard APIs for admin and timesheet logs. Contributed to homepage and landing page backend, ensuring clean user flow. 🔹 RPA (Robotic Process Automation) Learned RPA fundamentals, identified manual processes, and automated them. Designed efficient workflows for better task management and reduced manual effort. 🔹 QA & Testing Performed manual and automated testing using Selenium and Cucumber BDD. Wrote test cases and scenarios, ensuring smooth application performance. 🔹 Data Engineering Used Kafka and PySpark to process real-time wind turbine data. Stored cleaned data in Delta Tables to enable structured access and analytics. 🛠 Tech Stack: Python, Go, React.js, Swagger, AWS (EC2, S3, EKS, Cloud9), Docker, Kubernetes, PostgreSQL, Kafka, PySpark, Delta Lake, Selenium, Cucumber, Git, JSON