Bangalore, India · Open to Opportunities

Allam
Sai Krishna

"I turn raw data into decisions that matter."
currently:: [ ]

I'm a data professional at Capgemini with 3+ years navigating the full data stack — from writing SQL queries at midnight to shipping GenAI pipelines by morning. My path is deliberate: Analyst → Scientist. Every project I take on brings me one model, one insight, one experiment closer.

0+
Years in Data
Analytics → AI Engineering
0%
Reporting Automated
Real project impact
0+
Certifications
AWS, Azure, IBM & more
0+
ML Projects Shipped
From NLP to credit risk
// 01 — About

Signal, not Noise

My data story starts in 2018 when I chose to specialize in Data Science within Computer Engineering at Presidency University. That decision wasn't accidental — I was already obsessed with the idea that numbers, if listened to carefully, could narrate the truth about any system.

"I didn't choose data. Data chose me — through every dashboard that revealed something no one expected."

At Capgemini, I've worn many hats: Operations Analyst, Data Visualization Engineer, AI Automation Builder, and CreditLens Implementation Lead. Each role deepened my conviction — data is only powerful when it drives action.

My next chapter is Data Science: building models that don't just describe what happened, but predict what's coming. I'm actively sharpening my skills in deep learning, statistical modeling, and advanced GenAI pipelines — because the analyst in me wants to become the scientist.

Outside work, I'm the kind of person who reads ML papers for fun, experiments with LangChain on weekends, and finds genuine excitement in a well-tuned confusion matrix.

🔭
Curiosity-Driven Analysis
I don't stop at the dashboard. I dig into why metrics behave the way they do — uncovering root causes, not just correlations.
⚙️
End-to-End Ownership
From raw SQL extraction to polished Power BI report — I own the full pipeline. Messy data doesn't scare me; incomplete thinking does.
🤖
AI-Native Mindset
I integrated Claude AI + MCP into a real automation workflow before it became mainstream. I think in systems, not just scripts.
📢
Storytelling with Data
The best insight is worthless if it can't be communicated. I translate complex findings into decisions non-technical stakeholders can act on.
// 02 — Experience

The Journey

Every role has been a deliberate step — not just climbing a ladder, but expanding the radius of what I can do with data. Here's how the story unfolds.

AI Newsletter Automation Engineer
Oct – Dec 2025
Associate Consultant · Capgemini — Flagship AI Project
  • Architected an end-to-end GenAI newsletter pipeline using Model Context Protocol (MCP) + Claude AI — the kind of system that used to require an entire editorial team.
  • Integrated 7 live data sources: Web Search, ArXiv, GitHub, Product Hunt, Twitter (X), Gmail, and Local File System — all orchestrated intelligently by AI.
  • Slashed content creation time from hours → under 10 minutes with a Research → Editing → Designing multi-phase pipeline.
  • Demonstrated that agentic AI automation is production-ready — not just a proof of concept.
10 min content cycle 7 integrated tools 3-phase AI pipeline
MCPClaude AILangChainGenAIAutomationAgentic AI
CreditLens Implementation Lead
Sep 2025 – Present
Associate Consultant · Capgemini — Banking & FinTech Domain
  • Led full lifecycle implementations of Moody's CreditLens for multiple banking clients — credit risk workflows, compliance automation, and reporting.
  • Wrote complex PostgreSQL scripts that extracted, transformed, and delivered data exactly as each client's risk models required.
  • Customized risk scoring models, credit memos, approval workflows — aligning a powerful platform to unique client policies.
  • Became the go-to for data validation and reconciliation across cross-functional delivery teams.
PostgreSQLCreditLensCredit RiskFinTechData Validation
Data Analyst — Moody's Analytics
Dec 2023 – May 2024
Associate Consultant · Capgemini
  • Deep-dived into large enterprise datasets — surfaced patterns that led to a 15% efficiency gain and 10% error rate drop in operations.
  • Built interactive Power BI + Python dashboards that cut report turnaround by 30% — decision-makers got answers the same day, not the next week.
  • Automated batch operations using Control-M, improving on-time job execution by 15%.
  • Developed Excel KPI dashboards that improved real-time monitoring by 20% across teams.
15% efficiency gain 30% faster reports 10% fewer errors
PythonPower BISQLControl-MEDAExcel
Operations Analytics Engineer
Oct 2022 – May 2023
Associate Consultant · Capgemini — First role, biggest impact
  • Built reporting automation with SQL + Python + Flask that eliminated 80% of manual effort — what took hours now ran in minutes.
  • Created revenue, project, and resource dashboards that improved forecasting accuracy by 20%.
  • Optimized multi-source data workflows and cut project delays by 15% through smarter data pipelines.
  • Learned early: automation is empathy — every manual process saved is someone's Friday evening returned to them.
80% effort reduced 50% faster analysis 20% better forecasts
FlaskPythonSQLAutomationReporting
// 03 — Projects

Built from Scratch

01
GenAI · MCP
AI-Powered Newsletter Generator
A fully autonomous newsletter pipeline that researches AI trends, writes, designs, and delivers — completely hands-free. Built with Claude AI and Model Context Protocol, integrating 7 live tools from ArXiv to Gmail. This project redefined what "automation" means in content pipelines.
⚡ Hours → 10 minutes per newsletter
Claude AIMCPLangChainPythonGmail APIArXiv
02
NLP · Retrieval
Chat With Document
Upload any document, ask it anything. Built an interactive semantic retrieval system where LLMs don't just answer — they understand context across long documents. Used LangChain for chunking, embeddings, and vector search. A clean Streamlit UI made it accessible to non-technical users.
📈 25% improvement in query efficiency
PythonLangChainStreamlitLLMsVector Search
03
ML · PySpark
Credit Card Defaulter Prediction
Large-scale credit risk modelling using PySpark MLlib — designed for datasets too large for pandas. Applied feature engineering and statistical preprocessing to distinguish likely defaulters from safe borrowers. Benchmarked Logistic Regression, Decision Trees, and Random Forests with cross-validation.
🎯 18% model performance boost via feature engineering
PySparkMLlibRandom ForestFeature Eng.Cross-Validation
04
NLP · Sentiment
Sentiment Analysis Engine
End-to-end NLP pipeline for real-world text classification. Built from scratch — tokenization, stopword removal, TF-IDF vectorization, and model training. Evaluated across multiple classifiers. Solid foundation that later evolved into my understanding of transformer-based models.
PythonNLTKScikit-learnTF-IDFText Classification
05
ML · Fraud
Transaction Fraud Detection
Team-led project to catch fraudulent transactions before they slip through. Managed the complete pipeline: cleaning imbalanced data, engineering behavioural features, training and evaluating classifiers. First project where I understood that real ML is 80% data, 20% model.
PythonScikit-learnImbalanced DataFeature Eng.Team Lead
🔬
Next Project Loading...
Currently building something with Deep Learning + Time Series forecasting
// 04 — Skills & Learning

The Stack

Core Data & Analysis
Python
92%
SQL
90%
Pandas / NumPy
88%
Power BI
82%
Machine Learning & AI
Scikit-learn
80%
LangChain
78%
NLP
75%
GenAI / MCP
72%
Big Data & Cloud
PySpark
74%
PostgreSQL
82%
AWS
65%
Azure
62%
Visualization & Dev Tools
Matplotlib/Seaborn
84%
Flask / Streamlit
76%
Git
78%
Currently Upskilling
On The Horizon
Active learning stack — updated 2025
Deep Learning & Neural Nets 55%
PyTorch fundamentals, CNNs, RNNs
Advanced GenAI / RAG Pipelines 70%
RAG, vector DBs, agentic workflows
Statistical Modeling & Math 60%
Bayesian stats, hypothesis testing
Cloud ML (SageMaker / Azure ML) 40%
Model deployment & MLOps pipelines
Time Series Forecasting 45%
ARIMA, Prophet, LSTM for sequences
// 05 — Certifications

Proof of Work

Every certification here was earned with intent — not just for a badge, but to close a specific knowledge gap. AWS and Azure prove cloud readiness. IBM and DataCamp prove hands-on ML. CreditLens proves domain depth.

🏦
Moody's CreditLens Certification
Moody's Analytics · Credit Risk
☁️
AWS Cloud Practitioner
Amazon Web Services · CLF-02
🔷
Azure Fundamentals
Microsoft · AZ-900
🐍
Data Analysis with Python
IBM
📊
Supervised Learning with Scikit-Learn
DataCamp
🗄️
SQL for Data Science
UC Davis · Coursera
🔁
MLflow in Action: Master MLOps
Online · MLOps Track
🤖
LangChain Crash Course
Build OpenAI LLM Powered Apps
// 06 — Contact

Let's make
data matter

I'm actively looking for my next challenge — ideally somewhere that values both analytical rigor and creative problem-solving. If you're building something interesting in data, ML, or AI, I want to hear about it.

Actively Looking
What I'm looking for
Data Analyst / Senior Analyst roles
Junior Data Scientist positions
ML Engineer / AI Engineer opportunities
Analytics Engineering roles
GenAI Application Developer positions
Start a Conversation →