Hello, I'm

NIKHIL VARMA KANDULA

The data was always there. Someone had to make it useful.

Open to Opportunities · Worldwide
Nikhilvarma Kandula, Data Analyst
0
Years Experience
0
Tech Debt ↓
0
Users Served
Python· SQL· BigQuery· Power BI· Tableau· Apache Spark· Machine Learning· LLM Orchestration· GPT-4· Google Cloud· AWS· ETL Pipelines· Predictive Modelling· Looker Studio· Pandas· Python· SQL· BigQuery· Power BI· Tableau· Apache Spark· Machine Learning· LLM Orchestration· GPT-4· Google Cloud· AWS· ETL Pipelines· Predictive Modelling· Looker Studio· Pandas·
Agile / Scrum· Deep Learning· Firebase· Data Modelling· NLP· Spatiotemporal Analysis· REST APIs· Microservices· React.js· Statistical Analysis· SDLC· Business Intelligence· Agile / Scrum· Deep Learning· Firebase· Data Modelling· NLP· Spatiotemporal Analysis· REST APIs· Microservices· React.js· Statistical Analysis· SDLC· Business Intelligence·

About me

I'm Nikhilvarma — data engineer and analyst. I spent 1.5 years inside Fintech rebuilding pipelines that were costing teams hours every day. 500+ people now make faster decisions because of that work.

I'm studying for an M.Sc. in Big Data & Business Analytics at FOM Hochschule — still working on real problems while building the academic layer on top.

The best data work is invisible. The pipeline nobody notices because it never breaks. The dashboard that makes the answer obvious. That's what I build.

Right now I'm looking for the right role — somewhere the work is real and the learning keeps going.

EnglishC1 · Professional
DeutschA2 → B1 · Learning
TeluguNative
HindiNative

"Speed is a feature. Accuracy is non-negotiable. Both together — that's the standard."

— Nikhilvarma

Skills & Tools

The full stack of tools I use to turn raw data into real decisions.

32+
Tools & Platforms
4
Skill Categories
4
Cloud Platforms
5+
AI / LLM Tools

Data Languages & Tools

PythonSQLRDAXPandasNumPyBigQueryApache SparkFirebasePower BITableauLooker StudioText Mining

Analytics & ML

scikit-learnspaCyTensorFlowRandom ForestK-MeansMeta-Learner StackingTF-IDFSpatiotemporal AnalysisStatistical AnalysisETL PipelinesPredictive Modelling

AI & Modern Tech

GPT-4Claude APILlama 2/3LangChainHugging FacePrompt EngineeringLLM OrchestrationRESTful APIsMicroservicesAPI Security

Process & Methodology

Agile / ScrumAgile MethodologiesSDLCCode ReviewSprint PlanningDigital TransformationStorytelling with DataData-First CultureCross-functional LeadershipMentoring

Where I've made an impact

Every number below is real. Every result was earned — not estimated.

MicroIntech
Fintech · Remote · 🇺🇸 USA · Oct 2024 – Dec 2025
1.5 years
Data Specialist
40%
Tech Debt ↓
500+
Concurrent Users
15h+
Eng. Hours Saved / wk
25%
Data Latency ↓
01Spearheaded Digital Transformation

Migrated legacy monolithic financial systems to API-agnostic microservices — cutting BI reporting time by 50% and accumulated tech debt by 40%.

02Architected High-Performance Data Pipelines

Event-driven Python data pipelines serving 500+ concurrent users with 100% data integrity — zero missed financial transactions.

03AI & LLM Orchestration at Scale

RESTful APIs integrated with GPT-4 & Llama for automated financial audit — reduced manual review effort by 40% and saved 15h+ per week per analyst.

04Infrastructure Optimization

Firebase Firestore optimisation for high-frequency transaction monitoring — reduced data query latency by 25%.

05Leadership & Data-First Culture

Systematic code review process + mentoring for a team of 5+ engineers — production bug rate dropped 15% over 6 months.

PythonPandasNumPyGPT-4Llama 2/3LangChainFirebaseRESTful APIsMicroservicesAgile

Currently seeking working student / junior analyst role in Germany.
Available immediately · Student visa · 20h/week

Get in touch →

Built from scratch. Proven in results.

Real problems. Real data. Real outcomes — including a peer-reviewed publication, a Google Data Analytics capstone, 3,200 German tech postings analysed with NLP, and 2.91M flights modelled for delay risk.

6 Projects
5.5M Largest dataset (rows)
3,200 Job postings (NLP)
2.91M Flights analysed
40% Tech debt reduced
1 Peer-reviewed paper
01
★ Peer-Reviewed · IRJMETS · March 2025

Rainfall Estimation via Heterogeneous Data Fusion & ML

Traditional rain gauges create dangerous blind spots. This pipeline fuses Ground Sensors + Weather Radars + Meteosat satellite imagery into one signal via a Hierarchical Probabilistic Ensemble Classifier — outperforming the industry-standard Kriging with External Drift baseline.

Ground Sensors
Weather Radars
Meteosat
HPEC · POD 0.58
POD = 0.58 outperforms Kriging with External Drift (KED) — industry baseline
Python Random Forest Meta-Learner Stacking Spatiotemporal Analysis Meteosat
02
NLP + Clustering · FOM M.Sc. · 2026

German Job Market Intelligence Corpus

4,183 raw postings (StepStone + Indeed DE + LinkedIn) → 3,200 validated records → TF-IDF + spaCy EntityRuler NER → K-Means k=4 (silhouette 0.61). Result: Python + SQL co-occur in 65% of all postings — the empirical minimum viable skill pair for the German market.

Top skills by posting frequency
Python
84%
SQL
78%
Excel
59%
Power BI
50%
Azure
40%
3,200
Validated Postings
156
Skills Extracted
88.4%
NER Precision
0.61
Silhouette (k=4)
PythonspaCyTF-IDFscikit-learnK-MeansPCABeautifulSoup
03
🎓 Google Analytics Capstone · May 2026

Cyclistic Bike-Share: Casual → Member Conversion

5.5M real Divvy trip records — too large for spreadsheets, so: GCS → BigQuery → CTEs → window functions → Tableau. The headline finding: casual riders are NOT failed members. They are a different behavioural category. 5% conversion = ~97,500 new annual members.

Casual vs Member — key behavioural differences
Avg ride duration 19.2 min vs 11.8 min
Peak day Saturday vs Thursday
Seasonal drop −93% vs −75%
5.5M
Trip Records
1.63×
Casual Ride Length
93%
Seasonal Drop
3
Recommendations
BigQuerySQL CTEsWindow FunctionsTableau PublicGoogle Cloud StorageR · ggplot2
04
MicroIntech · Lead Developer · Oct 2024 – Dec 2025

Fintech Real Estate Intelligence System

Rebuilt a monolithic fintech platform into an event-driven, cloud-native data system serving 500+ concurrent users with financial-grade reliability. LangChain + GPT-4/Llama pipeline collapsed a 15h/week manual audit to 3h/week.

Partner PDFs & Portals
500+ User Events
LangChain · GPT-4 · Llama
PDF extract → ledger cross-check
Firebase Firestore
Sub-second monitoring
Power BI · RESTful APIs · Microservices
500+
Concurrent Users
40%
Tech Debt ↓
15h+
Saved / Week
25%
Latency ↓
PythonFirebase FirestoreLangChainGPT-4MicroservicesEvent-DrivenRESTful APIs
05
Statistical Framework · FOM M.Sc. Group Project · 2026

Quantifying Data Quality — Statistical Framework & Return-Rate Analysis

A four-dimension scoring framework (Completeness · Consistency · Accuracy · Timeliness) producing a single composite Data Quality Index — validated on UCI Air Quality (9,357 records). Composite DQI: 0.840. The only failing dimension: Timeliness (0.626), correctly identifying a 20-year-old dataset as unsuitable for real-time monitoring.

Data Quality Index — dimension scores
Completeness
0.930 PASS
Consistency
0.978 PASS
Accuracy
0.824 PASS
Timeliness
0.626 FAIL
Composite DQI 0.840
PythonpandasscipystatsmodelsUCI Air QualityHypothesis TestingJupyter
06
Predictive Analytics · BTS Data · 2019–2023

U.S. Flight Delay Analysis — Predictive Risk Modelling

2.91M BTS domestic flights (2019–2023). Late aircraft alone accounts for 41.1% of all delay minutes — more than weather, NAS, carrier, and security combined. Logistic Regression + Decision Tree predict delay probability from pre-flight schedule data alone (accuracy: 59.2%).

Delay cause breakdown (% of total delay minutes)
Late Aircraft
41.1%
Carrier
27.8%
NAS
18.3%
Weather
10.1%
Security
2.7%
2.91M
Flights Analysed
41.1%
Late Aircraft Share
17.5%
Overall Delay Rate
59.2%
Model Accuracy
Python 3.10pandasscikit-learnLogistic RegressionDecision TreeBTS DataJupyter

Education & Certifications

From First Class engineering to a Master's in Big Data in Germany — every step has been intentional.

  1. Oct 2025 — Aug 2027 · Germany · In Progress

    M.Sc. Big Data & Business Analytics

    FOM University of Applied Sciences

    Applying 1.5+ years of industry experience to academic research in Spatiotemporal Analytics, AI-driven automation, and advanced BI systems.

    Power BITableauBig Data EngineeringAI Automation
    Data & AI Big Data AnalyticsBig Data ArchitectureNLP & Unstructured Data AnalysisAdvanced ML for BusinessEnd-to-End ML Pipeline Project
    Engineering & Systems Applied Programming (Python & Data Engineering)Big Data Architecture & InfrastructureInformation Security & Data Privacy
    Analytics & Strategy Business AnalyticsSpatiotemporal Data AnalyticsBusiness Intelligence & Reporting SystemsData-Driven Decision MakingQuantitative Analysis & Statistics
  2. Feb 2026 · Completed · Credential JO1A2NXM2RU9

    Google Data Analytics Professional Certificate

    Google · Coursera

    Capstone: 5.5M real bike-share records · BigQuery + Tableau · 3 data-backed marketing recommendations delivered.

    SQLBigQueryTableauRData Cleaning
  3. 2020 — 2024 · Hyderabad, India · First Class with Distinction

    B.Tech in Computer Science & Engineering

    Malla Reddy College of Engineering

    Software Engineering · System Design · AI & Data foundations. Published peer-reviewed research in final year.

    Software EngineeringSystem DesignAI Foundations
Google · Coursera
Google Data Analytics Professional ⭐
Mar 2026
SAP
SAP Certified Associate — Data Analyst · SAP Analytics Cloud
In Progress · Expected Jul 2026
FOM University
Big Data Engineering & Architecture
2025–27
FOM University
Analysis of Semi- & Unstructured Data
2025–27
FOM University
Spatiotemporal Data Analytics
2025–27
FOM University
Business Intelligence & Reporting Systems
2025–27
FOM University
Advanced Machine Learning for Business
2025–27
FOM University
API Security & Data Privacy Compliance
2025–27
FOM University
Data-Driven Decision Making
2025–27
MicroIntech (on-the-job)
LLM Orchestration & Prompt Engineering
2024–25

Peer-Reviewed Research

Peer-Reviewed Publication · IRJMETS · March 2025

A Machine Learning Approach for Rainfall Estimation Integrating Heterogeneous Data Sources

K. Nikhil Varma · IRJMETS · Vol. 7 · Issue 3 · March 2025 · ID b3d0de1ee3008bdcbdae1ccb72560041
★ POD score of 0.58 — outperformed industry-standard KED baseline. Impact: flood early-warning · landslide risk prediction.
View Published Paper →

Open to Work

Let's work together· Data Analyst· Business Analyst· People Analyst· Marketing Analyst· Open to Germany· Open to India· Making the complicated simple· Data Analyst· BI Analyst· Junior Data Engineer· Open to Germany·
🇩🇪

Germany

Germany · Europe · Full-time · Working Student · On-site / Hybrid / Remote
Data AnalystBI AnalystData EngineerAnalytics EngineerBusiness AnalystReporting AnalystJunior Data ScientistJunior ML Engineer
🇮🇳

India

India · Full-time · Open to Relocate · Remote
Data AnalystData EngineerBI AnalystAnalytics EngineerBusiness AnalystJunior ML Engineer

Get in touch

Let's build
something meaningful.

If something here looked interesting — that's reason enough to reach out. I reply within 24 hours. Every time.

Available for new roles · Immediately