S&P Global
Senior Data Scientist
ASHUTOSH MISHRA / AI ENGINEER
Intelligence, built
for the real world.
I build AI that moves from possibility to production. Seven years of turning data into GenAI and machine learning systems used by millions.
Currently at S&P GlobalSenior Data Scientist · AI Lead
SCROLL TO EXPLORE01 — 07
01 / EXPERIENCE
AI in production.
Impact at scale.
Building intelligent systems across enterprise AI, personalisation, and language understanding.
Wipro
AI Tech Lead, GenAI Practice
Promoted from Data Scientist
Led enterprise AI and agentic systems delivery, including root-cause analysis for infrastructure diagnostics.
- GenAI engineers led
- 15
- Lower mean time to recovery
- 15%
More about the work at Wipro
- Fine-tuned and deployed domain-specific small language models on secure, GPU-based on-premises infrastructure.
- Partnered with client stakeholders across 6+ enterprise engagements, from solution scoping through production deployment.
- Defined architecture, evaluation criteria, and rollout plans for enterprise AI and infrastructure diagnostic solutions.
- Led a team of 15 GenAI developers building enterprise AI applications.
- Designed and developed GenAI applications for infrastructure root-cause analysis, reducing mean time to recovery by 15%.
- Fine-tuned large language models to generate task-specific recommendations, and trained small language models for root-cause analysis.
- Benchmarked language-model performance across hardware platforms.
Let's Cooee
Founding Engineer and Data Scientist
Built and scaled core AI and data products as an early founding engineer, including real-time purchase-intent systems.
- E-commerce sessions per day
- 11M+
- Data-processing efficiency gain
- 60%+
More about the work at Let’s Cooee
- Designed an event-driven architecture that improved data-processing efficiency by more than 60%.
- Developed an LLM-powered analytics system using tool calling over PostgreSQL, contributing to a 4%+ conversion uplift.
- Automated model training, evaluation, deployment, monitoring, and retraining pipelines.
- Built hyper-personalisation systems using genetic algorithms to generate notification image payloads and vary their text and call-to-action buttons.
- Improved user retention by 4.5% through personalised notifications driven by reinforcement learning.
- Reduced notification payload generation time by 10% through optimised database queries.
- Explored customer behaviour through recency, frequency, and monetary value (RFM) analysis to inform personalisation.
Riverus
Associate Data Scientist
Built machine learning systems for large-scale legal corpora, achieving 90% precision in extracting high-value analytics.
More about the work at Riverus
- Improved fuzzy text-matching performance by 25% for quoted-text comparisons across legal datasets.
- Developed and deployed an ensemble of transformer and supervised learning models supporting one-class classification for 160K+ users.
- Extracted high-value analytics from legal corpora, including common provisions, clauses, and structured data points from contracts.
- Built robust machine-learning pipelines and concurrent API endpoints, improving in-house API speed by 10%.
- Accelerated text-based CAPTCHA recognition using object-detection models.
- Built a transformer-based case summarisation tool to organise the chronology of legal proceedings.
- Developed a first-person reasoning model to identify speaker polarity.
- Built a contextual sentence-similarity model using natural-language understanding techniques.
02 / PRACTICE
Curiosity meets
engineering.
My work connects machine learning research with the systems that make it useful.
Intelligence
Generative AI, LLMs, NLP, deep learning, reinforcement learning, graph networks
Engineering
Python, SQL / NoSQL, cloud infrastructure, Docker, CI/CD, scalable ML pipelines
Research
Uncertainty quantification, Bayesian methods, time series, model benchmarking
FOUNDATIONS
IGIT
B.Tech, Computer Science · 2015–2019
Honours in Artificial Intelligence and Real Time Systems
Early work & internships
KG LLP Aug — Sep 2018
Data Science Intern · Financial time series, deep learning, and uncertainty quantification.
TINO IQ Jun — Aug 2018
Machine Learning Intern · Financial modelling and stock price prediction.
Intel Indexer LLC Feb — May 2018
Project Intern · Economic prediction, dynamic time warping, and Bayesian methods.
03 / SKILLS
Tools of the trade.
From agent design and model serving to the data and infrastructure behind production AI.
Languages
Python, JavaScript, SQL, Rust
Generative AI
LLMs, AI agents, RAG, MCP, tool calling, LLM fine-tuning, Hugging Face, LangChain, Google ADK, FastMCP
ML & inference
PyTorch, machine learning, deep learning, JAX, Scikit-learn, vLLM, TGI, model serving, Unsloth
Backend & cloud
FastAPI, Node.js, AWS, Azure, Google Cloud, Docker, CI/CD, MLOps, serverless deployment, Nginx
Data
PostgreSQL, MongoDB, ClickHouse, Snowflake, ChromaDB, vector databases, knowledge graphs, NoSQL
04 / ACHIEVEMENTS
Milestones along the way.
Recognition, open-source contributions, and work shared with the AI community.
- 2026 · HACKATHON WINNER
Hugging Face Build Small
Rune Goblin: an AI dungeon crawler where hand-drawn runes become spells. Read the build story ↗
- 2025 · TEAM RECOGNITION
S&P Global Innovation Award
Member of the award-winning team.
- 2020 · INDIVIDUAL RECOGNITION
Riverus Rookie of the Year
- 2017 · HACKATHON WINNER
Smart India Hackathon
Winner at the inaugural nationwide edition.
- 2026 · OPEN-SOURCE AMBASSADOR
OpenBMB
- OPEN SOURCE · CONTRIBUTOR
AI models & datasets
Synthetic datasets and fine-tuned models published on Hugging Face.
05 / WRITING
Notes from
the exploration.
The latest posts from the blog, on machine learning, research, and practical tools.
Machine Learning
Zephyr: Exploring Shape and Airflow with AI
Machine Learning
What I Learned Benchmarking Jev Against Classical Machine Learning
Hackathons
Rune Goblin: My Build Small Hackathon Winner
Deep Learning
GPU Server SetUp for DeepLearning
06 / CONNECT
Good things start
with a conversation.
Have an interesting problem in AI, an idea to explore, or just want to say hello?
prog.mishra@gmail.com ↗