// Engineer. Researcher. Zero Mediocrity.
Fast builder. Sharp thinker. Zero tolerance for mediocrity. I fine-tune LLMs, ship RAG pipelines, and wire real-time APIs — all the way to production, with tests and CI/CD. The future is being assembled — 92.4% accuracy and counting.
// 01. about
I'm Sanskruti Varade — a full-stack developer and AI/ML engineer with hands-on delivery across three domains: LLM-powered applications, cybersecurity tooling, and enterprise web systems.
I ship end-to-end — from RAG pipelines and NLP classifiers to multi-API threat dashboards with real-time geo-visualization. Published researcher in Generative AI for Cybersecurity at RTCSA 2026, currently completing MCS at K.K. Wagh College, Nashik.
One principle drives everything: build things that work in the real world, not just in notebooks. Available for opportunities from mid-2026.
// 02. skills
// 03. experience
// 04. projects
Fine-tuned DistilBERT on 55K IMDB reviews achieving 92.4% accuracy and 0.979 AUC-ROC. Async FastAPI with Redis caching, rate limiting, and CSV batch inference (500 rows/request). 30-test Pytest suite, Dockerized, GitHub Actions CI/CD — end-to-end LLM fine-tuning to live production endpoint.
Production RAG pipeline — multi-doc Q&A with source attribution, confidence scoring, 5-turn memory, FAISS vector store. Deployed live on Streamlit Cloud.
WebSocket multi-room chat with @cipher AI assistant (Groq Llama 3.3), live analytics dashboard, JWT auth, and Neon Cloud PostgreSQL via Prisma.
// 05. research
An adaptive honeypot system powered by Generative AI and LLMs — creating dynamic, believable decoy environments that lure and study real attackers. Significantly improves threat intelligence over static honeypots.
// 06. education
// 07. certifications
// 08. interview prep
Resources I've built and compiled to prepare for FAANG-level interviews — free for anyone to use.
STAR framework answers to the most common behavioral questions — leadership, ownership, conflict, failure, and why this company.
Pattern-first walkthroughs — not just solutions but the full thinking process from brute force to optimal with Python code.
REST APIs, databases, caching strategies, and scalability patterns — explained the way engineers actually think about tradeoffs.
// 09. blog
I write about Cybersecurity, AI/ML, Python, and software engineering — topics I actually build, not just read about.
How I used LLMs to make decoy systems that actually think back. From my RTCSA 2026 research paper.
→ Read · ~6 min Cybersecurity · APIs · VisualizationBuilding a dashboard that turns raw threat data into something you can actually see and act on.
→ Read · ~5 min AI/ML · Python · LLMs · RAGHow Retrieval-Augmented Generation turns a static PDF into a system you can have a conversation with.
→ Read · ~6 min DSA · Python · Interview PrepNot just solutions — the reasoning process behind them, from brute force to optimal.
→ Read · ~7 min Python · Beginners · Clean CodeSmall habits and built-in features that make your Python cleaner, faster, and more Pythonic.
→ Read · ~6 min// 10. contact
Open to internships, full-time roles, collaborations, and interesting conversations. Available from mid-2026.