Tejaswini Mediga

MCA graduate with 1+ years of hands-on experience building and evaluating machine learning models (Scikit-learn, AWS SageMaker) and multi-agent RAG systems (LangGraph, ChromaDB), with additional experience in prompt engineering, LLM output evaluation, and large-scale AI training data annotation for NVIDIA's autonomous driving pipeline (via Randstad Digital). Built RESTful APIs and real-time dashboards at JPMorgan Chase & Co. Skilled in Python, SQL, LangChain, LangGraph, and applied ML; seeking to grow into an AI/ML Engineer role building reliable, production-grade AI systems.

Biz Tech Analyst – Prompt Generation (AI) (2026-03 – Present)

• Designed and refined prompts for LLMs (OpenAI APIs, Claude) to generate accurate, contextually relevant outputs, improving instruction-following and model output quality.
• Evaluated AI-generated responses against quality benchmarks; identified gaps and optimized prompt structures to improve instruction-following.
• Implemented RAG pipelines using vector databases (e.g., ChromaDB) to enhance LLM knowledge retrieval and provide grounded, accurate responses for complex business use cases.
• Collaborated with cross-functional teams to translate use cases into effective prompt strategies, and documented prompt versions, test results, and best practices.

Process Executive – Data Annotation (AI/ML) at Randstad Digital Pvt. Ltd. (Client: NVIDIA) (2025-09 – 2026-02)

• Annotated LiDAR point clouds, traffic signals, pedestrians, and dynamic objects via the Human Loop platform, meeting NVIDIA's strict quality benchmarks for autonomous-vehicle training data.
• Delivered 6,720+ annotated frames/week over 7 months with an 85+ quality score, contributing training-ready data to NVIDIA's self-driving AI pipeline.
• Maintained consistent labeling standards across 7 annotation teams with zero escalation incidents.

Software Engineering Intern at JPMorgan Chase & Co. (2024-04 – 2024-09)

• Engineered 3 real-time dashboards using JPMorgan's Perspective library to visualize live trade data for internal use.
• Designed and integrated RESTful services to facilitate seamless communication between frontend dashboard components and backend trade data systems.
• Accelerated developer onboarding by configuring Git & Node.js setup, cutting setup time from 1 hour to 25 minutes.
• Resolved 10+ frontend bugs in a legacy codebase, reducing reload failures by 40% and improving page load speed.
• Streamlined team workflows by documenting setup steps and eliminating Git repo blockers, reducing support queries by 50%.

Data Analytics and Visualization Intern at Accenture (Forage – Virtual Internship) (2024-01 – 2024-04)

• Analyzed 7 cross-platform social datasets in Excel to identify content patterns, traffic sources, and audience drop-off points.
• Built 3 campaign-level Excel dashboards with pivot charts, reducing weekly reporting time by 40%.
• Delivered a 10-slide executive summary deck with KPIs and engagement insights, tailored for non-technical stakeholders.
• Recommended 4 content optimization strategies, supporting a projected 15% uplift in engagement based on trend analysis.

AI/ML Virtual Intern at EduSkills (2023-09 – 2023-11)

• Developed supervised ML models using AWS SageMaker, achieving up to 82% accuracy on structured and image datasets.
• Applied AWS Rekognition to preprocess and classify 2,000+ images for object detection and facial analysis tasks.
• Tuned model parameters and evaluated performance using confusion matrix and precision-recall metrics.
• Deployed end-to-end ML pipelines in a simulated cloud environment, showcasing scalability in real-world scenarios.
• Integrated MLOps workflows to track model lifecycle metrics and ensure deployment reliability within AWS SageMaker environments.

Master of Computer Applications (MCA) in Information Technology – Anurag University (2024-05)

Bachelor of Science in Computer Applications (B.Sc.) – Annie Besant College for Women (2022-07)