Rohank Khatri
Computer Science undergraduate with hands-on experience building and shipping API-integrated systems in Python – including LLM-backed retrieval pipelines, webhook-driven automation, and third-party API integrations (Whats App Business API, vector databases). Comfortable working across the stack from data pipelines to deployed, user-facing products.
Currently deepening backend fundamentals (relational databases, framework-based API development) and looking to apply strong Python and systems-integration skills in a production backend environment.
Experience
AI Automation & Analytics Intern
Aurika.ai
Remote
Built and maintained webhook-driven automation workflows in Python, integrating multiple third-party APIs and LLM services into a single pipeline. Designed structured data-processing logic to convert unstructured inputs into consistent, reusable outputs for downstream systems.
Tracked workflow performance and reliability, identifying and fixing integration failures across API boundaries. Collaborated asynchronously with a cross-functional team, maintaining clear documentation for handoffs and iteration.
Education
Jaypee Institute of Information Technology (JIIT-128)
B.Tech
Skills
PythonJavaScriptSQLREST APIsWebhooksThird-party API integrationWhatsApp Business APILLM APIsn8n workflow automationVector databasesQdrantStructured data extractionStructured data processingBasic relational DB usageOpenAI APIOllamaPrompt engineeringRAG pipeline designGitGitHubVS CodeLinux basicsDjangoDjango REST FrameworkPostgre
Projects
RecoBase – AI Recommendation Engine
Python, FastAPI, PyTorch, Qdrant, Gemini
Built a full-stack, India-localized recommendation ecosystem combining multiple ML models (DeepFM for CTR prediction, Two-Tower for retrieval, GRU for sequential/session modeling) served via a FastAPI backend. Implemented a Gemini-powered RAG pipeline over a Qdrant vector store for natural-language product search alongside the ML-based recommendations.
Built a companion Chrome extension to capture real user click/behavior events and an analytics dashboard to monitor model performance (latency, confusion matrices).
Flasher – AI Flashcard SaaS
Next.js 15, Firebase, Gemini
Built and deployed a document-to-flashcard SaaS app: users upload PDFs/DOCX/images, Gemini extracts key concepts, and the app generates a spaced-repetition study deck. Implemented Google authentication with route-level middleware, and used Firestore + Firebase Storage for data and file persistence with custom security rules.
Designed a tiered (free/paid) product structure with usage-based limits, shipped as a production-style SaaS rather than a demo.
Yeeter
Next.js, yt-dlp
Built and deployed a lightweight video-download utility with a Next.js frontend and a yt-dlp-based extraction backend.
Open Source
16+ public repositories on GitHub
Spanning AI/LLM applications, automation pipelines, and web projects.
Additional Information
Languages
English (Fluent), Hindi (Native)
Availability
Immediate
Location
Noida/NCR, open to relocation