Machine Learning Engineer · Data Scientist · AI Engineer
Building intelligent systems from data to deployment — Berlin, Germany 🇩🇪
Results-driven ML Engineer with 1.5+ years of industry experience — a 7-month internship followed by a full-time role at Pinaca Technologies — building and shipping machine learning models for real-world audio intelligence systems. Proficient in statistical analysis, data visualization, predictive modeling and Python-based ML workflows.
Currently pursuing advanced studies in Data Science in Germany, and seeking a Werkstudent position to apply and expand that ML expertise in a real engineering environment.
| Speaker diarization, neural noise suppression (GAN, DeepFilterNet, DTLN) and classification models — built, evaluated and run in production. | Cleaning, feature engineering and statistical analysis in Python, communicated through Power BI dashboards non-technical stakeholders can read. | Containerised services with Docker on Linux, orchestrated with Kubernetes (K3s) on AWS EC2, shipped by a GitHub Actions pipeline that builds, publishes and rolls out on every push. |
Junior Machine Learning Engineer — Pinaca Technologies · Aug 2024 – Jul 2025 · Vijayawada, India
- Engineered an end-to-end speaker diarization module for VIGOR, a production audio intelligence platform — designing a pipeline with precision labeling across dynamic audio streams and delivering consistent segmentation accuracy on large-scale real-world audio.
- Architected and deployed multi-stage noise suppression using DeepFilterNet and a Deep Temporal Linear Network (DTLN), measurably improving audio clarity and production system performance.
- Implemented precise speaker identification integrated with vector embedding databases, reducing false positives and increasing detection reliability in live audio streams.
- Deployed and monitored containerised ML services with Docker on Linux, ensuring stable production operations.
Machine Learning Intern — Pinaca Technologies · Jan 2024 – Jul 2024 · Vijayawada, India
- Researched GAN-based audio enhancement models, improving noise elimination and output quality — impact sufficient to earn promotion to a full-time engineering role.
- Developed a binary audio classification model (clean vs. noisy) on open datasets from the University of Edinburgh repository, strengthening the quality assessment pipeline.
Earlier roles
Data Science Intern — Unified Mentor · Sep 2023 – Dec 2023 · Remote, India
- Cleaned, preprocessed and feature-engineered large datasets with Python (Pandas, NumPy) for actionable business insights.
- Built and evaluated classification and regression models (Scikit-learn) with measurable accuracy improvements over baseline.
- Designed interactive dashboards in Matplotlib, Seaborn and Power BI for non-technical audiences.
ML & AI Intern — ExpertsHub · Jan 2021 · Bangalore, India
- Designed and trained MLP models for handwritten digit image compression and decompression, achieving significant storage reduction while preserving visual fidelity.
| Project | What it does | Stack |
|---|---|---|
CloudNotes DevOps |
Full-stack notes app deployed end to end on AWS: multi-arch Docker build, published to Docker Hub by GitHub Actions, rolled out to a K3s cluster on EC2 via kubectl set image. |
Flask · MySQL · Docker · GitHub Actions · Kubernetes (K3s) · AWS EC2 |
Local RAG Chatbot AI |
Ask questions of your own PDFs. Chunking → HuggingFace embeddings → ChromaDB retrieval → Llama 3 under Ollama, so answers stay grounded in the document. Runs fully locally. | LangChain · ChromaDB · HuggingFace · Ollama · Llama 3 |
AI Assistant with Web Search AI |
Gives a local LLM live information: a question triggers a DuckDuckGo search, results flow through a LangChain prompt template, Llama 3 returns a concise grounded answer. | LangChain · Ollama · Llama 3 · Streamlit · DuckDuckGo |
Audio Classification ML |
Binary clean vs. noisy classifier — the quality gate ahead of the audio processing pipeline at Pinaca. Trained on open University of Edinburgh datasets. | TensorFlow · Signal Processing · Jupyter |
H1N1 Vaccine Prediction ML |
Heavily imbalanced survey data — SMOTE oversampling plus Precision/Recall/F1 evaluation instead of accuracy, improving minority-class prediction over the baseline. | Scikit-learn · SMOTE · Pandas |
HR Analytics Data |
Predicts employee promotion from HR data, paired with a Power BI layer that puts the drivers in front of non-technical stakeholders. | Scikit-learn · Pandas · Power BI |
Medicare Data Analysis Data |
TF-IDF vectorisation over free-text fields combined with EDA, correlation analysis and inferential statistics to surface actionable insights. | NLTK · TF-IDF · SciPy |
Also on GitHub
| Project | About | Stack |
|---|---|---|
| Superstore Sales Dashboard | Total sales, profit and quantity across the Superstore dataset. | Power BI |
| HR Analytics — Attrition | Attrition, job role and employee detail reporting. | Power BI |
| Los Angeles Airbnb EDA | Exploratory analysis of the LA Airbnb listings dataset. | Python · Pandas |
| Telecom Churn Prediction | Predicting churn from telecom customer data. | Python · Scikit-learn |
| Movie Recommendation System | A recommender built over a movie dataset. | Python · Pandas |
| House Price Prediction | Regression modelling over housing data. | Python · Scikit-learn |
Programming & Data
Machine Learning & AI
Generative AI
Data & Visualization
Cloud & DevOps
Databases
flowchart LR
A[Developer<br/>local Flask] --> B[GitHub<br/>push to main]
B --> C[GitHub Actions<br/>automated build]
C --> D[Docker<br/>multi-arch image]
D --> E[Docker Hub<br/>versioned artifacts]
E --> F[Kubernetes<br/>K3s rolling update]
F --> G[AWS EC2<br/>cluster compute]
G --> H[(MySQL<br/>application data)]
B.Tech, Electronics & Communication Engineering — UCE Narasaraopet, JNTUK · 2018 – 2022
IEEE Member · Organiser of College Technical Fest · NSS Volunteer
PG Certificate, Data Analytics & Machine Learning — Imarticus Learning, Hyderabad · Nov 2022 – Aug 2023
Top 10 finish in the Imarticus Learning Hackathon (2023)
Certifications & languages
Certifications
- PG Certificate in Data Analytics & Machine Learning — Imarticus Learning
- Machine Learning using Python — Certified
- Python Core — Sololearn
- Robotics & Embedded Systems Workshop — IIT Hyderabad
- Top 10, Imarticus Learning Hackathon (2023)
Languages
| Language | Level |
|---|---|
| Telugu | Native · C2 |
| Hindi | C1 |
| English | B2–C1 |
| German | A1 · actively learning |
- Studying Data Science in Germany, and open to Werkstudent roles.
- Going deeper on Generative AI, Agentic AI and LLM agent architectures.
- Happy to collaborate on Data Science / AI projects and Cloud & DevOps automation.