Priyamvada Priyadarshani

AI Intern

0+ yrs exp 72 · Very Good

About

No bio added yet

Skills & Expertise (41)

LLMs Advanced

8.0/10

Years Exp

Prompt Engineering Advanced

8.0/10

Years Exp

Python Intermediate

7.5/10

Years Exp

SQL Intermediate

7.0/10

Years Exp

PyTorch Intermediate

7.0/10

Years Exp

FastAPI Intermediate

6.0/10

Years Exp

Git Intermediate

6.0/10

Years Exp

Docker Intermediate

5.5/10

Years Exp

C++ Intermediate

5.0/10

Years Exp

JavaScript Intermediate

5.0/10

Years Exp

Typescript Intermediate

5.0/10

Years Exp

System Design Tailwind CSS Postgres Supabase Firebase MongoDb Streamlit GCP LINUX Vercel Data Structures & Algorithms DBMS OOP Operating systems React NextJS REST APIs PGvector SQLAlchemy AsyncIO Ollama Qwen DeepSeek ResNet50 Transformers Hugging Face Vector Search Semantic Search Embeddings RAG

Work Experience

AI Intern

ProCertus AI

Apr 2026 - Present

Worked on the Retrieval layer & Query pipeline architecture of an enterprise Text-to-SQL/RAG pipeline over PostgreSQL data and unstructured Google Drive documents, enabling natural-language querying across heterogeneous enterprise data. Built a synthetic internal procertus benchmark spanning 15+ PostgreSQL tables across 3DBs & 1500+ documents and 1,000+ ground-truth queries using LLMs, and Faker to evaluate and highlight procertus product use-cases. Built an evaluation pipeline for namespace/semantic object retrieval, RAG, and SQL generation across SpiderBench, BIRD, and custom ProCertus benchmarks; identified and fixed retrieval/generation gaps, re-engineered prompts and evaluation flows, increasing Recall@1 from 85% to 98% and overall accuracy avg from 30% to 80%. Investigated LLM failure modes including generation loops, malformed outputs, hallucinated context, and incomplete responses; implemented structured-output validation, retry/fallback mechanisms, persona-based semantic caching, and cache invalidation to improve system reliability. Enhanced debug logs and platform user’s experience through persona managed responses.

SDE1 & Associate Technical Instructor

Newton School of Technology

Jul 2024 - Dec 2024

Designed and refined project-based programming curricula, developing 50+ coding assignments and guiding 600+ CS & AI students through software projects, DSA, and technical problem-solving. Mentored student teams during hackathons and technical projects, providing engineering guidance and code-level feedback; achieved an average student feedback score of 4.7/5.