Prashant Mishra
AI ML Engineer
+91 7692812444 | [email protected] | linkedin.com/in/prashant-mishra-sde007
Professional Summary
AI/ML & Generative AI Engineer with 3+ years of experience building scalable ML systems and deploying production-
grade GenAI applications. Specialized in LLM fine-tuning, RAG pipelines, and agent-based workflows using state-of-the-
art open-source and proprietary models. Proven track record of leading end-to-end development — from backend APIs
to deployment and monitoring — with a focus on cost efficiency, safety, and real-world business impact. Contributor to
patent-pending GenAI applications, and mentor to 100+ associates.
Skills
• Languages & Libraries: Python, Java, JavaScript, SQL, Pandas, Numpy, Scikit-learn, Matplotlib,
Seaborn
• Frameworks & Tools: Django, FastAPI, Flask, React, Docker, Git, DVC, Jenkins, Selenium
• ML/DL: PyTorch, TensorFlow, Hugging Face, Transformers, BERT, LoRA, QLoRA, PEFT, GAN, NLP,
Computer vision, ETL, EDA, Pyspark
• GenAI : Langchain, LlamaIndex, LangGraph, CrewAI, AutoGen, PydanticAI, RAG, Prompt Engineering
• Cloud & Deployment: Azure OpenAI, AWS Bedrock, Google Vertex AI, vLLM, Ollama, OpenLLM
• Monitoring & Evaluation: MLflow, Comet Opik
• Databases: PostgreSQL, MongoDB, SQLite
• Soft Skills: Technical mentorship, Cross-functional teamwork, Fast ramp-up on new tech
Experience
Tata Consultancy Services May 2022 – Present
Machine Learning Engineer
1. Multi-Turn Interview Assistant:
• Led development of an agent-based full-stack GenAI app using LangChain, LangGraph, Azure OpenAI,
Django, and React.
• Simulated client interviews with dynamic follow-up questions, reducing training & evaluation effort by
60%.
• Adopted modular agent design for easy extension across new domains.
2. RAG-Based Document Chatbot:
• Built a scalable Retrieval-Augmented Generation (RAG) system using hybrid retrieval, self hosted vector
stores, and open-source LLMs.
• Improved retrieval precision and reduced hallucinations by 50–60% using context-aware chunking and
reranking.
• Ensured enterprise-level data privacy with secure access controls and PII-masking.
3. Centralized Guardrails for LLM Safety:
• Created a NLP-based guardrails platform with fine-tuned BERT based models, Presidio, and Django–
PostgreSQL backend.
• Enabled bidirectional filtering (user & model), improving LLM safety by 80–90% across 3+ POCs.
• Reduced redundant effort across teams by centralizing safety layers.
4. LLM Fine-Tuning & Optimization:
• Fine-tuned LLaMA 3.2 and Phi-3 models using LoRA, QLoRA, SFT, and PEFT on internal domain datasets.
• Transitioned organization from proprietary to open-source models, cutting inference costs and boosting
alignment.
5. Red Teaming & GenAI Safety Evaluation:
• Designed red-teaming protocols for GenAI apps to detect bias, toxicity, and injection vulnerabilities.
• Created custom test datasets and evaluation scripts; integrated feedback into safety layers pre-
deployment.
6. Model Monitoring and Evaluation:
• Conducted model monitoring and response logging for machine learning applications using evaluation
metrics such as BLEU, ROUGE, Levenshtein score, and F1 score, leveraging tools like MLflow and Opik.
• Implemented prompt versioning for comparative analysis of new prompts and features to drive continuous
improvement in model performance.
Python Developer
1. On-Premises and Cloud LLM Deployment:
Deployed open-source LLMs using inference frameworks such as vLLM, Ollama, and OpenLLM to support
internal GenAI POCs, enabling team-wide access and significantly reducing reliance on paid LLM services.
This infrastructure cut development and experimentation costs substantially, accelerating POC creation
while maintaining control over model usage.
2. AI Playground Application:
Developed a Django-based platform leveraging client-side Jupyter Notebooks for Python and ML
assessments. Deployed in production and actively used by associates, it reduces manual test creation time
by 60% through automated test case generation and features a custom Python package for efficient local
execution, ensuring secure user access with OAuth2 authentication.
Education
Master of Computer Applications July 2023 – July 2025
Chandigarh University
Bachelor of Science in Computer Science July 2019 – July 2022
APS University, Rewa, Madhya Pradesh
Achievements & Awards
1. Mentored 100+ associates in Machine Learning, Deep Learning and Generative AI, contributing to upskilling
and capability building across teams.
2. Contributed to the development of two Generative AI applications submitted for patent filing.
3. Recipient of 4 TCS On-the-Spot Awards for exceptional contributions to AI/ML projects and innovation
initiatives.
4. Delivered 3+ production-grade AI applications in under 3 years, demonstrating technical proficiency and agility
Certifications
• Microsoft Certified Azure AI Engineer Associate
• AWS Certified Machine Learning Specialty
