Amit Kumar Jaiswal
AI/ML engineer
Report
professional-data verificationHiring assessment
6 role-fits · 0 risksLLM judgment on best-fit roles, bounded by available evidence.
Also a strong fit
Since Oct 2021 he has focused on RAG and GenAI at MIPT, and since Jan 2023 has delivered freelance RAG/LLM solutions using FAISS, LangChain/LangGraph, FastAPI, and Docker. His part-time Academic Mentor role at MIPT (since Oct 2023) indicates capability to teach foundational ML/AI topics and guide student projects.
Photo gallery
Card images from confirmed sources where the subject is featured. Click any tile to open the source page. Not face- verified — some thumbnails may be article hero shots that show another person.
LinkedIn + Resume summary
LLM-distilled, not verbatimAI/ML engineer and PhD candidate specializing in RAG, agentic LLM systems, and secure GenAI
Mauritius · Email: ameejais0999@gmail.com · Email: amitjaiswallpu@yahoo.in · LinkedIn: https://linkedin.com/in/amit-kumar-jaiswal-5bbb59156 · WhatsApp: +91 7042387265 · Telegram: @ameejais
AI/ML engineer with 5+ years building production RAG pipelines, agentic LLM workflows, and secure GenAI systems. Blends applied research at MIPT with freelance and consulting delivery, including fintech components at BranchX. Focuses on reliability, safety, and measurable impact using vector search, prompt defenses, and scalable FastAPI deployments.
- Reduced hallucinations by over 40% in a multilingual YouTube RAG system via hybrid retrieval and cross‑encoder reranking (LangChain/LangGraph, Chroma).
- Benchmarked LLMs under adversarial prompt injection; observed ~68% average leak rate and built Unicode normalization and obfuscation defenses.
- Delivered production-aligned RAG assistant and agentic workflows at BranchX, exposing reliable inference and retrieval via FastAPI services.
- Proposed adaptive cumulative entropy thresholding for IoT DDoS detection; recognized with 1st place at MSU’s 2024 AI in Education conference.
- Published on anomaly detection approaches (IJCSIS, 2024) and IoT DDoS entropy method (Cybersecurity Issues Journal, 2025).
- Built a Gemini-based multi-agent research assistant generating citation-backed reports end-to-end in under a minute.
Mentored undergrads on neural nets, NLP, and RL; reviewed Python ML code and guided capstone projects from design to evaluation.
Delivered production RAG pipelines, multilingual assistants, and LLM security hardening using LangChain, FAISS, and FastAPI for diverse clients.
Built secure RAG architectures, fine-tuned domain LLMs, and studied adversarial robustness for assistants and cybersecurity applications.
Implemented a RAG knowledge assistant and agentic workflows; shipped FastAPI inference services and tuned prompts/models for reliable fintech use.
Completed hands-on labs in Wireshark analysis, WordPress brute-force testing, and session hijacking using Kali and Metasploit.