Prateek Pathak

AI Researcher

Ahmednagar, Maharashtra, India2 yrs 4 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Reduced chatbot response times by 40%.
Optimized HR operations, cutting query volumes by 30%.
Enhanced storage solutions, reducing data retrieval latency by 30%.

Stackforce AI infers this person is a Data Scientist specializing in AI and scalable data solutions.

Contact

Skills

Core Skills

Generative AiNlpScalable Ml Pipelines

Other Skills

Apache DruidRayRedisLangchainPythonPyTorchSQLTableaumatplotlibcloud solutionsdata pipelinesRAG systemssemantic routingsystem designWeb Development

About

Hi, I’m Prateek Pathak, a passionate Data Scientist at Jio Platforms Limited with a proven track record of engineering innovative AI systems and scalable machine learning pipelines. From deploying Retrieval-Augmented Generation (RAG) chatbots to creating robust hybrid cloud data solutions, I thrive at the intersection of data, technology, and innovation. In my current role, I’ve developed cutting-edge AI systems that deliver measurable impact. My highlights include: • Reducing chatbot response times by 40% increasing user satisfaction significantly. • Optimizing HR operations, cutting monthly query volumes by 30% through innovative RAG systems. • Enhancing storage solutions to cut data retrieval latency by 30%. With expertise in Generative AI, NLP, and scalable ML pipelines, I have hands-on experience with advanced tools like Apache Druid, Ray, Redis, and Langchain. I also bring deep technical proficiency in Python, PyTorch, and SQL, complemented by skills in visualization tools like Tableau, matplotlib.

Experience

2 yrs 4 mos

Total Experience

2 yrs 2 mos

Average Tenure

2 mos

Current Experience

Citi

Gen AI Developer

Feb 2026 – Present · 2 mos · Hybrid

Generative AINLPscalable ML pipelinesApache DruidRayRedis+6

Jio platforms limited (jpl)

Data Scientist

Dec 2023 – Feb 2026 · 2 yrs 2 mos · Navi Mumbai · On-site

Engineered a Retrieval-Augmented Generation (RAG) chatbot leveraging Llama 3.3/GPT-4o , enhancing query resolution by 25% through semantic caching.
Partnered with cross-functional teams to deploy a hybrid cloud-based storage solution using Apache Druid, reducing data retrieval latency by 30%.
Built scalable data ingestion pipelines using Ray (2.37.0) for seamless vector database management.
Optimized HR query resolution processes by introducing Redis-backed RAG pipelines, resulting in a 30% decrease in monthly query volumes.
Designed and implemented a semantic routing system, enabling the team to handle 30% more requests daily while improving speed by 50%.
Leveraged tools like Apache Airflow and MLflow for efficient workflow monitoring and machine learning lifecycle management.