N

Nirbhay Tiwari

CTO

Delhi, India5 yrs 2 mos experience

Key Highlights

  • Expert in building production-grade AI systems.
  • Led significant projects in Generative AI and LLMs.
  • Proven track record in data-driven decision making.
Stackforce AI infers this person is a Generative AI Engineer with expertise in AI systems and data analytics.

Contact

Skills

Core Skills

Generative AiLarge Language Models (llm)Data AnalyticsMachine LearningPredictive ModelingData AnalysisStatistics

Other Skills

AI AgentsArtificial Intelligence (AI)LangChainData ScienceIT Business AnalysisNeural Networkshugging faceFine TuningRlhfDatabasesStatistical Data AnalysisData VisualizationMicrosoft Power BISQLAnalytical Skills

About

Hi, I’m nirbhay tiwari — a Generative AI Engineer focused on building real-world, production-grade AI systems. I work at the intersection of LLMs, RAG architectures, and intelligent system design, where the goal is not just to build AI models, but to make them reliable, scalable, and usable in real-world scenarios. Currently at Birlasoft, I’m working on a RAG-based medical AI application built on large-scale proprietary and confidential datasets. My work involves designing end-to-end retrieval pipelines, improving context understanding, building agentic architecture, writing effective system prompts across application and ensuring the system delivers accurate and explainable responses in a highly sensitive domain. Before this, at Turing, I contributed to Google’s Gemini project, where I worked on large-scale LLM improvement workflows. I was part of a team delivering 2,500+ high-quality Colab notebooks focused on enhancing model understanding, improving response quality, and reducing error rates. I also designed 500+ real-world and synthetic API scenarios for LLM training, and worked on structured prompt testing that improved model responsiveness and consistency. In addition, I’ve developed and deployed AI/ML solutions for EV startups, supply chain startups, where I worked on tasks like text classification and intelligent data extraction, achieving high accuracy and delivering measurable business impact. What I Do I Build end-to-end RAG systems (chunking strategies, embeddings, retrieval optimization, bot response evaluation pipeline) Design agentic workflows for complex problem-solving Work on LLM evaluation, prompt engineering, and model reliability Handle challenges like hallucination reduction, domain adaptation, and response accuracy Tech Stack Python | FastAPI | SQL Generative AI | RAG | Agentic AI | NLP LLM Fine-tuning | RLHF | SFT | Prompt Engineering Data Analysis | Visualization Beyond Work I also enjoy leading a team of ai enthusiasts and have led a team of data scientists. I’m always curious about pushing AI systems beyond prototypes — into robust, production-ready solutions that actually solve problems at scale.

Experience

5 yrs 2 mos
Total Experience
--
Average Tenure
--
Current Experience

Birlasoft

Tech lead Gen AI Development

Jun 2025Present · 1 yr · India · Hybrid

AI AgentsArtificial Intelligence (AI)Generative AILarge Language Models (LLM)LangChain

Turing

Data Scientist

Apr 2024May 2025 · 1 yr 1 mo · Cary, North Carolina, United States · Remote

  • Google Gemini Generative AI Project:
  • Generalized AI Agent Development: Designed and implemented autonomous multi-agent pipelines using LangGraph and AutoGen. These agents could independently trigger actions such as deleting Slack messages for a specific user/date, drafting emails, fetching real-time external data, and summarizing documents. Ensured complete end-to-end task execution with strong assertion checks to validate tool behavior and response reliability.
  • Reinforcement Learning from Human Feedback (RLHF): Led RLHF training initiatives by collecting and curating human preference data, designing reward functions, and training reward models to guide LLMs toward responses aligned with real-world user intent, safety policies, and preferred style. This improved ranking capabilities and ensured model-generated outputs met judgmental criteria.
  • LLM Response Evaluation: Performed qualitative and quantitative analysis of model-generated responses, measuring correctness, coherence, style, and safety. Evaluated LLM outputs across use cases to ensure responses are contextually accurate, user-aligned, and factually sound.
  • SXS Evaluations & Stress Testing: Conducted Side-by-Side (SXS) evaluations to compare output quality across different model versions. Designed stress test prompts to uncover behavioral edge cases and identify emerging failure modes or bias patterns, contributing to continuous improvement.
  • Super-Fine Tuning on Domain-Specific Data: Fine-tuned models on curated domain-specific corpora (medical, legal, financial, etc.) using tailored instruction templates. This significantly improved LLM performance in high-accuracy-required scenarios and domain compliance.
  • Reasoning Capability Enhancement: Applied techniques like chain-of-thought prompting, few-shot examples, and ablation testing to elevate the model’s multi-step reasoning skills. This improved the ability to handle complex instructions, logical breakdowns, and structured query resolution.
Data ScienceIT Business AnalysisNeural NetworksGenerative AIArtificial Intelligence (AI)hugging face+4

Nimbuspost

Data Scientist

Sep 2023Mar 2024 · 6 mos · Gurugram, Haryana, India · On-site

  • Responsibilities & Contribution:-
  • Inventory Health Analysis: Conduct ABC analysis, inventory turnover, and stock-out analysis to optimize inventory levels, reduce carrying costs, and ensure products' availability while minimizing excess stock.
  • Customer Behavior Analysis: Utilize customer order history and feedback data to understand purchasing patterns, preferences, and trends, contributing insights for personalized marketing strategies or product placement.
  • Cancellation Trend Analysis: Analyze historical data to identify patterns and trends related to order cancellations. Investigate reasons such as stock unavailability, shipping delays, or customer dissatisfaction.
  • Customer Segmentation Analysis: Segment customers based on order cancellation behavior, demographics, purchase history, and interaction patterns to understand which groups are more prone to cancellations and why.
  • Feedback and Sentiment Analysis: Analyze customer feedback, reviews, and support interactions to understand the underlying reasons for dissatisfaction leading to cancellations and gauge sentiment trends.
  • Analysis Tools Used:
  • Python - Pandas
  • Mysql
  • Power Bi
  • Adv Excel - Power Query, Macros Automation
  • NLP
  • Machine Learning
Data AnalyticsDatabasesStatistical Data AnalysisData VisualizationMicrosoft Power BISQL+3

Chargeup

Data scientist

Nov 2022Sep 2023 · 10 mos · Delhi, India · On-site

  • Responsibilities & Contribution-:
  • Built a predictive model for driver churn prediction, enabling proactive measures to retain customers and reduce customer attrition. The model accurately identified customers at risk of churning, allowing targeted retention strategies to be implemented, leading to a 15% decrease in customer churn rate.
  • Conducted in-depth analysis of battery usage patterns and identified opportunities to reduce the number of idle batteries in the swapping stations. By implementing data-driven strategies, such as optimizing battery allocation algorithms and adjusting pricing plans, idle battery numbers were reduced by 25%, resulting in improved operational efficiency and cost savings.
  • Utilized statistical analysis on IOT data received from batteries to identify key factors influencing battery performance, such as temperature, charging patterns, and usage frequency. This analysis provided valuable insights for improving battery maintenance procedures, resulting in a 30% increase in battery lifespan and reducing maintenance costs.
  • Collaborated with cross-functional teams, Conducted market research and analyzed demographic and geographical data to identify potential target cities for new battery swapping stations. These insights facilitated informed decision-making and successful expansion into three new cities, contributing to a 40% increase in revenue.
  • Build data-driven fraud detection system that compares the data received from swapping stations with data received from IOT device and calculate the swap & revenue gap and help identify the possible dealers(swapping station) doing fraud in swaps and revenue.
  • Created comprehensive reports and data visualizations using Power BI to effectively communicate analysis findings and actionable insights to stakeholders.
  • Analysis Tool:
  • Python Library: Pandas, NumPy, Statistics, Matplotlib, Seaborn
  • Excel & Google Sheets
  • Power BI & Tableau
  • Data Modeling
  • MySQL
  • Machine Learning
Statistical Data AnalysisData VisualizationMicrosoft Power BIMySQLAnalytical SkillsPython (Programming Language)+5

So infotech - india

Data Analyst

Jun 2022Nov 2022 · 5 mos · Delhi, India · On-site

  • Key Responsibilities:
  • worked as an intern and gained better understanding of extracting data using SQL queries, cleaning, transforming data moreover finding actionable insights for organisation by doing explanatory indepth statical data analysis, building powerful dashboards on Power BI, forecasting based on past data.
StatisticsDatabasesStatistical Data AnalysisData VisualizationMicrosoft Power BISQL+14

Meta

Partner

Aug 2021Jul 2022 · 11 mos · Ireland · Remote

  • I was working here as a contractual content creation partner where my responsibility was to create content for meta/facebook and gain retention on content as per contract.

Bunk tribe holidays pvt ltd

Web Developer, Digital Marketer

Jan 2020Dec 2020 · 11 mos · Noida, Uttar Pradesh, India

  • Key responsibilities:
  • Responsible for Maintaining Functionality Of Website Both Front End Back End
  • Maintaining Database Of Company, Performing CRUD Operations on Data, Updating Website on regularly As Per Company Requirements.
  • Responding to Website Visitors Queries & Redirecting Them To Sales Professionals, Running & monitoring Facebook ads, doing search engine marketing, Writing Blogs for website, Keeping Track Of Users Activity/Action On Website, Collection Of Customers Data From Contact Forms And Providing To Business Development Managers.

Ea sports

Technical Support Engineer

Aug 2018Jun 2019 · 10 mos · Noida Area, India

  • Key responsibilities:
  • I Was Responsible For Fixing Glitches, Bugs Related To Newly Released Software/Products
  • Helping Users With Their Technical Issues, Testing of New released Games Glitches, Handling Customer Queries Efficiently.
  • Providing Fixes To Their Technical Problems With Easy To Understand Troubleshooting Steps.

Education

Indian Institute of Technology, Guwahati

Certification — Data Science

Jan 2025Nov 2025

CHANDIGARH UNIVERSITY

Bachelor of Technology - BTech — Mechanical Engineering

Jun 2014May 2018

Central Board of Secondary Education

Higher Secondary Board — Science

Jul 2013Apr 2014

Stackforce found 100+ more professionals with Generative Ai & Large Language Models (llm)

Explore similar profiles based on matching skills and experience