Work

Experience and skills.

Where I have worked, what I have shipped and the skills I bring to a team.

Download resume Get in touch

Experience

Roles and independent work

Jun 2026 — Present

AI / ML Engineer

BestPeers

Automated first-line IT and access support for a global enterprise client across several business applications: an AI agent triages each ticket, resolves what it can and escalates the rest to a human agent.
Built hybrid search over Milvus with embeddings to match each ticket to the right resolution template, falling back to a prompted Azure OpenAI response when nothing similar exists.
Designed a three-tier fallback (template, then LLM, then human agent) in Python services, so every ticket has a path to resolution.

Jun 2022 — Apr 2024

Data Scientist Intern / Jr. Data Scientist

MoreYeahs Technologies · Indore

Shipped Sentry-B, a real-time edge surveillance system running at 25+ FPS on a Jetson Nano: person detection (YOLOv3), tracking (DeepSORT), face extraction (MTCNN) and identity verification from live CCTV, optimized with TensorRT.
Moved operations at a streaming-TV client from reactive to proactive monitoring: Dataiku ETL over logs from 10 critical machines into an AWS S3 data lake, an Isolation Forest model that predicts failures, and MLflow tracking with automated alerting dashboards.
Delivered OCR-based financial document pipelines and LLM-powered NLP applications as FastAPI microservices on GCP, tuning extraction accuracy directly with clients.

2026

Independent open-source projects

github.com/sanidhya-ai-ml

DocuMind RAG, a RAG API with hybrid retrieval, reranking and evaluation
Research Agent Crew and Mega AI, multi-agent LLM systems
An MCP server and job-search automation agents

Project highlights

Selected projects in more detail

GenAI clinical trial source data verification platform

LangGraphRAGMCP serversOCRNERFastAPICeleryPostgreSQL

Replaced a fully manual clinical data verification process for a life sciences client with a LangGraph multi-agent pipeline that maps source documents to EDC records in CDISC ODM XML.
Built a RAG, OCR and NER layer that extracts clinical facts from PDFs and links each one to database records with full evidence traceability, so verification is auditable and source-backed.
Added a mismatch-detection engine and MCP servers that expose tools to agents as modular components, delivered as async FastAPI services with Celery queues for large document batches.
Owned the project solo, from requirements to production.

AI-powered resume optimization engine

LangChainChromaDBFastAPIPydanticOpenAI APIEmbeddings

A RAG-based tool that compares resumes with job descriptions semantically, surfacing genuine skill gaps instead of keyword mismatches.
A custom ATS scoring engine that jointly scores keyword alignment, skill relevance and experience context.

Skills

What I work with

GenAI and LLMs

RAG, hybrid search
LangChain, LangGraph
Multi-agent systems, MCP servers
Azure OpenAI, OpenAI API
RAGAS, prompt engineering
Fine-tuning (LoRA, PEFT)

Vector databases

Milvus
ChromaDB
Pinecone, FAISS
Embeddings

ML and deep learning

PyTorch, TensorFlow
Scikit-learn, Transformers
NLP, OCR, NER
Pandas, NumPy

Computer vision

YOLOv3, DeepSORT, MTCNN
OpenCV, TensorRT
Edge deployment (Jetson Nano)

Backend and APIs

Python (async, Pydantic)
FastAPI, REST, microservices
Celery, Redis
PostgreSQL, SQL, NoSQL

MLOps and cloud

Azure AI, GCP, AWS S3
Docker, Kubernetes
MLflow, Apache Beam, Dataiku
Git

Education and learning

Where I studied

Education

M.Tech in Data ScienceDAVV University, School of Data Science & Forecasting · Indore2024 — 2026
B.Tech in Information TechnologyMedi-Caps University · Indore2019 — 2023

Certifications

The MCP Course: Unit 1, Fundamentals of MCPHugging FaceApr 2026
Microsoft Azure AI Essentials: Workloads and ML on AzureLinkedIn LearningApr 2026

Open to

New roles and select freelance projects.

The quickest way to reach me is email.

Get in touch