SAMARTH PRATAP SINGH
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 LangGraph & RAG.
AI/ML Engineer with expertise in LangGraph, RAG, and Agent-based systems. Currently gaining practical experience as an Artificial Intelligence Applications Development Intern, specializing in document processing workflows, vision extraction pipelines, and robust output validation. Proficient in building full-stack RAG applications, scalable multi-agent research engines, and generative AI dialogue summarizers, leveraging skills across Python, FastAPI, Docker, and various ML/DL frameworks. Seeking to apply strong problem-solving and implementation skills in an innovative AI/ML environment.
VIT Bhopal University
B.Tech · Computer Science and Engineering
August 1, 2023 – June 30, 2027
AmberFlux EdgeAI Private Limited
Artificial Intelligence Applications Development Intern
May 1, 2026 – Present
Hyderābād, Telangana, India
Argus - Scalable Multi-Agent Research Engine
February 1, 2026 – February 1, 2026
Built a LangGraph supervisor-based multi-agent pipeline with 5 specialist agents (planner, researcher, critic, writer, supervisor) using LLM-driven Command(goto) routing to autonomously synthesize cited markdown reports - reducing manual research time from hours to 30-90 seconds. Designed async submit → poll → fetch job flow with SQLite persistence and LangGraph checkpointing for failure recovery; traced every LLM call end-to-end in LangSmith for observability. Containerized with Docker, deployed on Render with a health-check endpoint, and integrated Tavily web search, ArXiv, and Wikipedia as agent tools.
View ProjectDoCopilot (RAG Document Q&A)
December 1, 2025 – December 1, 2025
Built a production full-stack RAG app: upload PDFs/TXT, index with Qdrant hybrid search (BM25 + dense vectors), and retrieve cited answers; deployed with a FastAPI backend and Next.js frontend. Improved retrieval quality via RRF fusion + cross-encoder reranking; added guardrails (prompt-injection detection, PII redaction, source-grounding checks) to harden the pipeline against adversarial inputs. Ran a 40-query LLM-as-Judge ablation study reporting 89.2% avg correctness, 90.5% relevance, 100% source rate, and 2.86s avg latency - demonstrating measurable retrieval quality over baseline RAG.
View ProjectFLAN-T5 Dialogue Summarizer (GenAI)
October 1, 2025 – October 1, 2025
Deployed an interactive Gradio app on Hugging Face Spaces with configurable decoding (beam search, max length); published the fine-tuned model with a full model card and reproducible evaluation suite (ROUGE, BERTScore, METEOR, BLEU). Fine-tuned FLAN-T5-base with LoRA on SAMSum (14.7K dialogues), achieving 49.01 ROUGE-1, 72.25 BERTScore F1, and 42.51 METEOR. Implemented parameter-efficient training (LoRA r=16, α=32), updating only 2% of parameters with FP16 mixed precision - matching full fine-tuning ROUGE-1 at a fraction of the compute cost.
Deep Learning Specialization
deeplearning.ai / Coursera
January 1, 2025 – Present
Applied Machine Learning in Python
Coursera
January 1, 2025 – Present
LangChain for LLM Application Development
deeplearning.ai
January 1, 2025 – Present
Cultural Fit Analysis
The candidate's personal projects showcase a proactive and self-driven learning attitude, exploring diverse areas within AI/ML from RAG to multi-agent systems and fine-tuning. The detailed descriptions and quantitative results in projects like DoCopilot and FLAN-T5 demonstrate a commitment to quality and measurable impact. The current internship further solidifies their practical application of AI in an enterprise setting, indicating a good fit for a role requiring hands-on development and problem-solving.
Soft Skills & Operational Fit
The candidate demonstrates strong problem-solving skills through the implementation of complex AI systems and optimization techniques. The detailed project descriptions suggest a methodical approach to development and a focus on measurable outcomes. The use of observability tools like LangSmith indicates an understanding of operational best practices for AI applications.