HARSH SHARMA
Generative 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
Generative AI Engineer with 1+ years in LangChain & MLOps
Generative AI & LLM Engineer with 1.5+ years of hands-on experience building production-grade LangChain / LangGraph pipelines, Retrieval-Augmented Generation (RAG) systems, and MLOps infrastructure. Deep expertise in multi-stage RAG orchestration (hybrid retrieval, cross-encoder reranking, multi-turn memory), LLM tool-use and prompt engineering, and agentic workflow design with LangChain LCEL and LangGraph. Delivered measurable outcomes: +32% RAG precision@3 over a 500-query benchmark, +8 pp ROC-AUC uplift on imbalanced datasets, and 99.5% uptime on a 50K-record/day Databricks pipeline. Proficient across the full LLM delivery stack FastAPI microservices, Docker/Kubernetes, AWS SageMaker, and CI/CD with GitHub Actions.
NIT Delhi
B.Tech · Computer Engineering
August 1, 2020 – June 30, 2024
NIT Delhi
Data Analysis & NLP Intern
April 1, 2021 – May 1, 2021
Delhi, Delhi, India
AI Search Engine - Production RAG Retrieval System
June 23, 2026 – Present
• Architected a multi-stage RAG pipeline (embed→ retrieve → rerank → generate → filter → cache) delivering sub-second cached responses; latency validated with Locust load tests. • Implemented FAISS vector index + PostgreSQL metadata store for hybrid semantic + keyword retrieval; reranking results logged in eval/results.json for reproducibility. • Containerised 5 production microservices with Docker Compose (health checks, rate limiting, graceful shutdown); deployed on AWS EC2 with Nginx reverse proxy and Redis caching.
AI Resume Screening & Job Match System
June 23, 2026 – Present
• Built a PDF resume parser and NLP skill-extraction pipeline using spaCy NER and Sentence-Transformers to identify candidate skills and map them to a curated taxonomy. • Computed semantic similarity scores between resume and job-description embeddings via cosine similarity, generating a ranked match score and per-role skill-gap analysis. • Deployed an interactive Streamlit app and published a clean Kaggle notebook covering EDA, methodology, and evaluation metrics (precision, recall, F1) with inline visualisations.
AI Interview Coach
June 23, 2026 – Present
• Designed a multi-stage LangChain LCEL pipeline with four independent chains (Skill Extractor → Question Generator → Answer Evaluator → Feedback Synthesiser), enabling modular prompt iteration, per-stage unit testing, and clean separation of concerns. • Integrated GPT-40 to generate role-specific technical and behavioural interview questions from a parsed resume, then evaluate answers for communication clarity and technical depth - producing a scored improvement plan. • Persisted session state in PostgreSQL to allow users to resume coaching sessions, compare performance across attempts, and track skill-gap closure over time. • Deployed with Docker Compose (FastAPI backend + Streamlit frontend) with setup docs, screenshot walkthroughs, and a recorded demo in the README.
Distributed Real-Time Anomaly Detection System
June 23, 2026 – Present
• Benchmarked event-processing throughput up to 10,000 events/second using a Kafka + Redpanda setup with fault-tolerant consumer groups. • Improved anomaly detection precision by 18% by adding an online learning feedback loop, measured on a labelled holdout window after 24 hours of live retraining.
AI Developer Professional Certificate
IBM
January 1, 2024 – Present
Cultural Fit Analysis
The candidate's project portfolio is highly aligned with a Generative AI Engineer role, showcasing diverse applications of AI/ML, from RAG systems and NLP to real-time anomaly detection. The self-employed 'Generative AI & ML Engineer' role further emphasizes a proactive and self-driven approach, which can be a strong cultural fit for innovative teams. The breadth of technologies used across projects (AWS, Azure, Docker, Kubernetes, various ML frameworks) indicates adaptability and a willingness to explore different tools. The personal projects demonstrate initiative and a passion for the field.
Soft Skills & Operational Fit
The candidate demonstrates strong problem-solving skills through complex project architectures (e.g., multi-stage RAG pipelines, LangChain LCEL). The detailed project descriptions suggest a methodical approach to development, including testing, evaluation, and deployment considerations. The focus on measurable outcomes indicates a results-oriented mindset. However, without direct interview data, assessing collaboration, stress handling, and adaptability is limited.