# Harsh Tiwari

Harsh Tiwari is a remote worker on Himalayas, Machine Learning Engineer at Uniblox.io, based in India.

Source: https://himalayas.app/@harshtiwari4

```text
Data Scientist/ AI-ML Engineer building production AI systems, conversational agents, and scalable ML pipelines.
```

| Detail | Value |
| --- | --- |
| Availability | Open to opportunities |
| Location | India |
| Authorized to work in | India, Nepal |
| Job categories | Data Scientist, Machine Learning Engineer, MLOps Engineer, Backend Engineer, Computer Vision Engineer, Search Data Scientist, AI Data Scientist, Data Science, Data Science Specialist, ML Data Scientist, Research Data Scientist, Data Science Engineer, Machine Learning Scientist, Data Scientist Analytics |
| Skills | Python, SQL, PostgreSQL, MySQL, BASH, JavaScript, S3, PyTorch, TensorFlow, Hugging Face, Lang Chain, OpenAI, Gemini, OpenCV, RAG, fastAPI, Flask, REST APIs, Celery, Redis, Async Flows, Microservices, Docker, Kubernetes, AWS Lambda, AWS Sagemaker, Git, EC2, GitHub Actions, CI CD, Pinecone, Faiss, ChromaDB, Neo4j, Pandas, NumPy, Spark, Matplotlib, Power BI, Tableau, Vector Databases, Semantic Search, Recommendation Systems, Real Time AI Inference, GitHub, Strapi, XGBoost, S3 Lambda, NodeJS |
| LinkedIn | https://https://www.linkedin.com/in/harxhy/ |

## What I'm looking for

```text
I’m looking for a team where I can own production AI end-to-end—building scalable pipelines, shipping conversational/NLP and real-time inference, and tightening MLOps with strong engineering practices.
```

## About

```text
I’m an Data Scientist/AI-ML Engineer who architects end-to-end production AI systems for enterprise clients, combining Python, AWS, Docker, and Kubernetes with full lifecycle ownership. I build Python microservices and RESTful APIs with FastAPI and ship conversational AI workflows using GPT-4 Turbo/4o/mini, LangChain, and vector databases like Pinecone and FAISS.

I focus on practical ML engineering: I develop data pipelines, ML inference modules, and async task queues (Celery, Redis) to reduce deployment friction—cutting deployment cycles by 25%. I also build end-to-end AI pipelines for NLP and model deployment with PyTorch, Hugging Face, and FastAPI, improving accuracy by 15% and efficiency by 22% through better ML infrastructure and CI/CD.

In addition to product work, I deliver results through applied research and contract engagements—benchmarking models from BERT and XGBoost to OpenAI and Gemini, and strengthening prediction accuracy by 5–10% with feature engineering and tuning. My medical imaging work used CNNs/Transformers in PyTorch to achieve 92% accuracy and included large-scale processing (10TB+) and graph-based modeling in Neo4j. I bring that same end-to-end, production-minded approach across RAG, semantic search, recommendations, and real-time inference.
```

## Experience

### Machine Learning Engineer

Uniblox.io. Apr 2026 - May 2026

```text
Designed, trained, and benchmarked ML and GenAI models (BERT, XGBoost, OpenAI GPT, Gemini) with SHAP explainability and iterative experimentation to improve prediction accuracy by 5–10%. Owned end-to-end MLOps and deployment using Python, AWS (S3, SageMaker), Docker, REST APIs, and model endpoints for real-time production inference.
```

Skills: Python, PyTorch, BERT, XGBoost, OpenAI-API, Gemini, SHAP, Feature-Engineering, AWS-Sagemaker, MLOps

### Associate Data Scientist

Emplay Analytics Pvt. Ltd.. Jan 2025 - May 2026

```text
Architected end-to-end production AI systems for enterprise clients, taking full lifecycle ownership. Built conversational AI services using FastAPI and GPT-4, and developed backend pipelines and async task queues (Celery/Redis), cutting deployment cycles by 25%.
```

Skills: Faiss, Lang-Chain, Python, AWS, GPT-4-APIs, Pinecone, Docker, Kubernetes, Celery-and-Redis

### Machine Learning Engineer

Astatine.ai. May 2024 - Jan 2025

```text
Built end-to-end AI pipelines for NLP and model deployment, improving accuracy by 15% and efficiency by 22%. Integrated ML services with Node.js/Strapi and delivered 10+ RAG/semantic search/recommendation features while processing 500GB+ datasets.
```

Skills: Lang-Chain, fastAPI, NodeJS, PyTorch, Hugging-Face, AWS, S3, GitHub-Actions-CI-CD, Strapi, ChromaDB, pandas-NumPy-Spark, Sagemaker

## Education

### M. S. Ramaiah University of Applied Sciences

Bachelor of Technology in Mathematics and Computing, Computer Science. 2021 - 2025

```text
Completed a B.Tech in Mathematics and Computing (Computer Science) at M. S. Ramaiah University of Applied Sciences from 2021 to 2025.
```

## Tech stack

- GitHub
- Kubernetes
- GitHub Actions
- NumPy
- Pandas
- MySQL
- PostgreSQL
- Gmail
- Google Analytics
- Neo4j
- OpenCV
- Redis
- Python
- TensorFlow
- PyTorch
- FastAPI
- Gemini
- Strapi
- Docker
- s3-lambda
- SQL
- XGBoost
- Hugging Face
- LangChain
- ChromaDB
- Pinecone
- Bash
- Faiss
- Task
- Remote

To contact Harsh, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [Akhil Salla](https://himalayas.app/@akhilsalla): Software Engineer, Nexquare
- [ranjith rondla](https://himalayas.app/@ranjithrondla): Data Scientist, Intellectuals AI Solutions Pvt Ltd
- [srikar kodakandla](https://himalayas.app/@srikarkodakandla): Machine Learning Engineer, Taffi
- [nishikanta nayak](https://himalayas.app/@nishikantanayak): Power BI & Python Developer, Epicminds Information Technology
- [DIVYA SONI](https://himalayas.app/@divyasoni1): Data Analyst in Research and Development, Force Motors
- [sravani bandela](https://himalayas.app/@SravaniBandela): Data Scientist, Anarix.ai
- [Vedit Agrawal](https://himalayas.app/@veditagrawal): Applied AI Intern, ScaleDux
- [Raveesh Sehrawat](https://himalayas.app/@raveeshsehrawat): ML & Full Stack Trainee, Centre for Development of Telematics (CDoT)
- [VAIBHAV SHELAR](https://himalayas.app/@vaibhavshelar): Machine Learning Engineer, Eucloid Data Solutions
- [Brijesh Kapadiya](https://himalayas.app/@brijeshkapadiya): Data Science Engineer, Potenza Global Solutions
