Charchit Sharma
Verified Expert in Engineering
ML Engineer and Developer
Varanasi, Uttar Pradesh, India
Toptal member since January 29, 2026
Charchit is a senior machine learning engineer specializing in building, optimizing, and deploying scalable artificial intelligence systems, with deep expertise in generative AI, large language models (LLMs), and computer vision. His focus is on transitioning complex models into high-throughput production environments, reducing inference latency, and building robust, automated data pipelines.
Portfolio
Experience
- PyTorch - 7 years
- Python - 7 years
- Natural Language Processing (NLP) - 7 years
- Computer Vision - 7 years
- Large Language Models (LLMs) - 5 years
- FastAPI - 5 years
- Open-source LLMs - 4 years
- Agentic AI - 3 years
Preferred Environment
Linux, PyTorch, vLLM, Open-source LLMs, Docker, Python, Hugging Face
The most amazing...
...thing I've built is a highly scalable deepfake detection solution that processed 20,000+ daily requests, using a CLIP-based architecture to achieve 95%+ AUC.
Work Experience
Senior ML Engineer
IDfy
- Built and productionized a CLIP-based deepfake detection system using parameter-efficient fine-tuning (PEFT), achieving an AUC greater than 95%.
- Deployed the solution in a production Video-KYC pipeline capable of processing 20,000+ daily requests for identity fraud prevention.
- Architected a retrieval-augmented generation (RAG) pipeline handling 1+ million requests per day, and optimized vector search (Qdrant) to reduce inference turnaround time by 84%.
- Engineered a context-aware document classification engine using Qwen3-4B and optimized the back-end inference using vLLM to ensure high-throughput processing and low latency.
- Developed an automated regulatory compliance platform utilizing LLaMA-3.1-8B, implementing custom parsers and multi-stage pipelines to reduce manual audit time by 90%.
Applied ML Engineer
Avataar.ai
- Engineered an automated data curation pipeline for diffusion-based 3D reconstruction models, applying semantic segmentation and clustering to automatically filter 66% of noisy data.
- Developed distributed evaluation pipelines and optimized data loaders, ultimately reducing evaluation times by 30%.
- Deployed and evaluated multiple diffusion-based 3D models across distributed AWS cloud infrastructure utilizing GPU clusters.
Computer Vision Engineer (Research Assistant)
IIIT Hyderabad
- Benchmarked and evaluated over 150 pretrained CNN and ViT models to test for model robustness under real-world data corruption.
- Contributed core engineering code to major open-source repositories, including the HuggingFace Diffusers library and Facebook's Py-IRT library.
- Published paper at the ICLR 2023 workshop: https://arxiv.org/abs/2409.04041.
- Scaled educational resources and technical infrastructure for an NPTEL Computer Vision course, supporting over 7,000 enrolled students.
Systems Engineer
Infosys
- Was part of the Apple Global Business Intelligence team; implemented a data pipeline for named-entity recognition using a pre-trained transformer model, enabling efficient extraction of relevant entities from unstructured text data.
- Performed data analysis and validation on upstream data sources to identify quality issues and anomalies, improving the reliability of entity extraction workflows in production.
- Went through the training program organized by Infosys Limited.
Experience
Inspect-AI | Automated Compliance Auditing Platform (DPDP Act)
AI Scheduling Assistant | LLM and Google Calendar Integration
Education
Bachelor's Degree in Computer Science
Rajasthan Technical University - Jaipur, Rajasthan, India
Certifications
ML Summer School
Cohere
Deep Learning Specialization
Coursera
Apply Generative Adversarial Networks (GANs)
Coursera
Skills
Libraries/APIs
PyTorch, Pandas, NumPy, Pydantic, vLLM, Google Calendar API
Tools
Claude, DeepSeek
Languages
Python
Platforms
Linux, Docker, Google Cloud Platform (GCP)
Frameworks
Flask
Other
Open-source LLMs, Hugging Face, Transformers, FastAPI, Parsers, Natural Language Processing (NLP), Computer Vision, Machine Learning Algorithms, Large Language Models (LLMs), Agentic AI, Agentic RAG Systems, Anthropic, Vector Databases, Machine Learning, Generative Adversarial Networks (GANs), Deep Learning, Computer Science, Diffusion Models, BERT
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring