ADITYA THAKUR
AI EngineerAI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI Engineer with less than a year in LLM Applications & RAG Pipelines
AI Engineer with hands-on experience building and deploying LLM-based applications, Retrieval-Augmented Generation (RAG) pipelines, and NLP solutions. Proficient in LangChain, Hugging Face Transformers, Python, and vector databases (FAISS, Pinecone). Demonstrated ability to architect end-to-end AI systems from data ingestion through model deployment, achieving 85-90%+ answer relevancy benchmarks. Adept at prompt engineering, agentic workflows, and optimizing system accuracy and scalability using Docker and CI/CD practices.
DAV University
B.Tech · Computer Science & Engineering
August 1, 2022 – June 30, 2026
S.D.S.V.M Talwara
Senior Secondary (12th Grade)
June 1, 2021 – May 31, 2022
S.D.S.V.M Talwara
Secondary (10th Grade)
June 1, 2018 – May 31, 2019
Coder Roots
Deep Learning & NLP Trainee
June 1, 2025 – July 31, 2025
India
07 Services
Data Science & Machine Learning Trainee
June 1, 2024 – September 30, 2024
India
PDF Q&A RAG Application
June 24, 2026 – Present
• Built an intelligent PDF Question-Answering system using LangChain, PyPDF, Hugging Face Transformers, and FAISS for document querying. • Developed complete pipeline: PDF text extraction, chunking, embedding creation using sentence-transformers, and similarity-based semantic retrieval. • Integrated open-source LLaMA models for accurate response generation; optimized retrieval via chunk-size tuning and metadata filtering. • Achieved 85-90% relevance accuracy on user test cases; deployed with a clean Streamlit-based UI on Hugging Face Spaces.
Indian Cybercrime Legal RAG Assistant
June 24, 2026 – Present
• Built a Retrieval-Augmented Generation (RAG) application for Indian cyber law research, enabling users to describe situations in plain English and receive AI-generated legal analysis grounded in the IT Act 2000, IPC, and BNS 2023. • Designed and curated a structured dataset of 153 real-world cybercrime scenarios across 12 categories (hacking, financial fraud, cyberstalking, data theft, etc.) with mapped legal sections, explanations, and punishments. • Engineered a semantic search pipeline using FAISS vector indexing and Sentence-Transformers (all-MiniLM-L6-v2) embeddings to retrieve the most relevant legal provisions for any user query with sub-second latency. • Implemented a LangChain-based RAG chain that retrieves context from the FAISS vector store, formats multi-field legal metadata, and generates grounded responses via Groq-hosted LLaMA-3.3-70B with temperature-controlled inference. • Developed a conversational Streamlit interface with chat history, configurable retrieval depth (top-k), example query shortcuts, source transparency (showing retrieved entries with full metadata), and sidebar API key management. • Ensured privacy-first architecture by running all embedding generation locally on-device, with no user data leaving the machine during indexing; only the final query is sent to the Groq LLM API for inference.
YouTube Q&A RAG Application
June 24, 2026 – Present
• Developed a YouTube-based Retrieval-Augmented Generation (RAG) application for transcript-based question answering using LangChain, Streamlit, and Hugging Face Transformers. • Implemented end-to-end pipeline: YouTube transcript extraction, text chunking, embedding generation via sentence-transformers, and semantic retrieval using FAISS vector database. • Integrated open-source LLMs for contextual answer generation; achieved 90%+ answer relevancy on evaluation queries. • Designed interactive Streamlit interface and deployed on Hugging Face Spaces for public access.
Cultural Fit Analysis
The candidate's portfolio of personal projects, particularly the Indian Cybercrime Legal RAG Assistant, shows a strong interest in applying AI to real-world, impactful problems. This indicates a proactive and problem-solving mindset. The breadth of skills across LLM frameworks, NLP, vector databases, and deployment tools suggests adaptability and a willingness to learn new technologies. The focus on open-source tools and models also aligns with a collaborative and community-driven culture. However, the candidate's experience level is very junior (still pursuing a B.Tech degree), which might require more mentorship and structured guidance in a senior team environment.
Soft Skills & Operational Fit
The candidate demonstrates strong initiative and a project-oriented approach, which aligns well with roles requiring self-starters. The focus on end-to-end application development, including UI and deployment, suggests a practical, results-oriented mindset. The privacy-first architecture in the legal RAG project indicates an awareness of ethical considerations and robust design. However, without direct interview data, assessing collaboration, stress handling, and communication clarity in a team setting is not possible.