Sidra Effendi

• Conducted data analysis for a cross-cultural online safety research with a global team, focusing on issues like cyberbullying and stalking. Consolidated and structured data for pattern recognition and integration. Managed virtual collaboration using Zoom and shared documentation via Google Workspace. Deployed surveys via Qualtrics, facilitated secure data transfer between UofM and partner universities, and identified patterns, which informed victim interview design. • Designed the system architecture for SuperAbbu, an AI-powered maternal health assistant leveraging LLMs, RAG, LangChain, Whisper for voice synthesis, and a Vector DB for contextual search, targeted at under-resourced healthcare environments. Presented a real-time conversational automation infrastructure using Retrieval-Augmented Generation and open-source AI tooling, focused on scalable impact in South Asia. • For AI security, contributed to PyRIT (Prompt Injection Risk Identification Toolkit) by integrating a prompt-generation training database to support LLM safety testing and jailbreaking resistance. • Managed remote collaboration using Slack, Zoom, and Google Workspace; coordinated stakeholders across time zones and led documentation efforts with Otter.ai and Google Notes. Scheduled and facilitated meetings; delivered detailed follow-ups and ensured alignment across cross-functional teams. Applied Excel (including PivotTable) for tracking research outputs and maintaining structured summaries of qualitative findings. • Co-authored research accepted at NNVAWI (Nursing Network on Violence Against Women International); additional paper under review at CSCW Journal.

• Led development of an NLP-powered Digital Curation Assistant for UN OCHA's ReliefWeb platform for humanitarian PDF management, resulting in a 13× increase in productivity. Gathered requirements from ReliefWeb’s CTO, Editor, & Program Manager, working on editorial workflows, business needs, & integration. • Managed a cross-functional team of engineering & data science, using Kanban. Personally built the MVP, NLP features & the entire search engine, transitioning to project lead: hiring, onboarding, mentoring, assigning tasks, & ensuring team continuity, while working on the next version of the product. Designed & architected a solution for AI integration into editorial workflows. Validated needs with stakeholders, focusing on development, empowering editors through automation & structured search. Ensured knowledge transfer to maintain project reliance. Used the ReliefWeb API for data access for model training & testing, ensuring the open-source non-profit project nature. • Developed & integrated custom ML models: a neural network for disaster detection & a bi-relevance model to classify documents across 17 UN Sustainable Development Goals (e.g. poverty, education, agriculture). Applied Named Entity Recognition (NER), Part-of-speech (POS) tagging, text filtering, Linear SVC, Facebook BART & built APIs with FastAPI, using SpaCy & PyTorch. • Engineered a domain-specific vertical search engine that outperformed Google & Bing for NGO content, boosting search accuracy by extracting structured metadata & returning page-level results. Scaled the dataset tenfold via a custom Bash-based scraping & filtering pipeline. • Redesigned backend API architecture (Django to Flask/FastAPI) to resolve memory issues using a two-step PDF processing flow with exponential backoff, increasing API success. Deployed via PythonAnywhere with a GitHub-hosted static frontend, utilized Google Docs, and suggested key UX features to enhance editor usability (e.g. visual extraction indicators).

• Provided design, development, and deployment of an IoT-based automation system for 2,000+ retail locations, for clients like Domino’s, Airtel, Vodafone, and Amul. Programmed edge devices, including Raspberry Pi, temperature control sensors, actuators, and contact sensors with value-based logic to control ACs, lighting, and ovens based on real-time conditions. Deployed updates remotely using MQTT for reliable message delivery in low-connectivity environments. Ensured secure device access via VPN and handled firmware updates and authentication. Responsible for development, testing, architecture, documentation, and coordination. • Designed and implemented Docker-based digital replicas (“device twins”) of Raspberry Pi-based edge devices synced with the physical hardware to enable remote monitoring, predictive maintenance, and real-time system visibility. Built a custom SDK for file management and Git-based syncing of code and configuration, allowing remote command execution and status reporting. Used InfluxDB, a time-series database, to ingest telemetry data, visualized device performance, and event logs through Grafana dashboards. Developed and tested in PyCharm, validated via unit tests and hardware, and managed deployments via Git workflows. • Participated in early-stage testing of the GNest protocol stack as part of a transition effort from IPv4 to IPv6, aimed at improving scalability and expanding the number of addressable IoT devices across custom hardware deployments.

• Built the full backend architecture and handled all server-side development, data storage, and testing for a photo-sharing app with individualized uploads and access. The app gained 100 customers organically and was selected for NASSCOM 10,000 Startups, DigitalOcean Hatch, and IBM Global Entrepreneur. • Worked with Postgres via a Flask-based cloud REST API. Created a serverless microservice on AWS Lambda, accelerating image-grid load and improving user experience. Implemented Google OAuth for authentication and leveraged AWS security best practices. • Worked on product design and GTM strategy, provided analysis, customer interviews, and investor pitches. Led UX and market research efforts to validate the concept and identify pain points. Handled UI/UX development using Figma, InVision, and Adobe Illustrator.

Courses: PP 639 - Introduction to Statistics and Data Analytics SI 618 - Data Manipulation and Analysis EECS 549 -Information Retrieval SI 671 - Data Mining EECS 548 - Information Visualization SI 630 - Natural Language Processing SI 670 - Applied Machine Learning SI 617 - Choice Architecture SI 563 - Game Theory

Courses: ● Theory of Computation ● Data Structures and Algorithms ● Operating System ● Database Management Systems ● Software Engineering ● Introduction to Logic