Open to AI engineer roles · internships now, full-time 2027 · remote or on-site

Abhinav Tarigoppula

AI/ML Engineer · Open-Source Contributor · IEEE Author

I build AI systems that hold up in production, and fix the open-source libraries they run on.

Final-year B.Tech CSE (AI/ML) at GITAM. Three AI internships including PwC, a fine-tuned vision-language model, a multi-agent system, and 50 bug fixes merged across 30 open-source orgs, including unsloth, pydantic, Black, asyncpg, redis-py, Celery and HuggingFace.

Now Fixing edge cases in JAX, dask and pyfakefs found by differential testing against NumPy, SciPy and the real OS.

Abhinav Tarigoppula
50
PRs merged upstream
30
open-source orgs
400k+
combined GitHub stars
3
AI internships

Counts from GitHub, updated November 2026

Worked at

  • PwC
  • AVAIntern
  • OneStop AI

Code merged into

  • unsloth
  • mem0
  • agno
  • Black
  • JAX (Google)
  • pydantic
  • marimo
  • HuggingFace
  • Snipe-IT
  • Prowler
  • redis-py
  • OGX
  • asyncpg
  • scapy
  • TagStudio
  • Xorbits
  • Celery
  • pymodbus
  • Galaxy
  • AMD GAIA
  • pyroute2
  • pyfakefs
  • Microsoft (semantic-link-labs)
  • Sklearn-genetic-opt
  • loky
  • fakeredis
  • Weaviate
  • Apache Maven
  • Mercari
  • palinode

About

Most AI demos work in a notebook and fall apart in production. I like the part in between: agents that call tools safely, retrieval that returns the right context, fine-tuned models that are measured honestly, and async code that doesn't freeze an event loop.

That last one turned into CodeHound, a static analyzer I wrote. The bugs it found were fixed and merged in unsloth, agno, pydantic, Black and Weaviate, and I have since had fixes merged in Celery, asyncpg, redis-py, scapy, pyfakefs, pymodbus and Apache Maven. Every one came with a regression test and went through the maintainers' review.

Alongside that I've shipped ML in three internships (PwC, AVAIntern, OneStop AI), had a paper accepted at IEEE ISED 2026, and placed Top 5 at the BITS Pilani AI/ML hackathon.

What I do

LLM, RAG & agents

4cooperating agents in AgentDesk

  • Multi-agent systems with validated tool calls
  • RAG that cites its sources
  • Prompt-injection and PII guardrails
  • Eval harnesses that catch regressions

Model training & fine-tuning

25→92%extraction validity after QLoRA fine-tuning

  • PyTorch, Transformers, PEFT / QLoRA
  • Vision-language models (Qwen2-VL)
  • Held-out and adversarial evaluation
  • Every result measured against a zero-shot baseline

Backend & full stack

−18%inference latency at OneStop AI

  • FastAPI services and REST APIs
  • Async Python that never blocks the loop
  • SQL, Docker and CI/CD
  • React front ends, schema to UI

Tech I use

Languages

  • Python
  • TypeScript
  • JavaScript
  • Java
  • SQL

AI / ML

  • PyTorch
  • Transformers
  • PEFT / QLoRA
  • unsloth
  • scikit-learn
  • OpenCV
  • Pandas

LLM

  • LangChain
  • OpenAI API
  • Claude API
  • MCP
  • RAG
  • FAISS
  • Tool calling
  • Evals

Engineering

  • FastAPI
  • asyncio
  • React
  • Node.js
  • Docker
  • AWS
  • GitHub Actions
  • pytest
  • Git

Experience

  1. PwC logo

    AI & Data Advisory Intern

    PwC · Launchpad Advisory Program

    Feb 2026 – Jul 2026Remote

    • Selected for PwC's flagship advisory program on GenAI, prompt engineering and data systems.
    • Built and optimised LLM prompt workflows for real enterprise use cases.
    • Worked with enterprise data architectures and AI-driven advisory solutions.
    • Reached Level 10 with 1500+ XP, top performance in the cohort.
    • GenAI
    • Prompt engineering
    • Data systems
  2. AVAIntern logo

    AI/ML Engineering Intern

    AVAIntern · AvaIntern Edutech

    May 2026 – Jun 2026On-site · Visakhapatnam

    • Built and evaluated 5+ ML models on 10,000+ record datasets, reaching up to 92% prediction accuracy through feature engineering and tuning.
    • Automated preprocessing and analysis with Python, Pandas and scikit-learn, cutting manual data preparation by 40%.
    • Ran EDA and model evaluation across 15+ features, improving model performance by 18%, and presented findings through visualisations.
    • Python
    • Pandas
    • scikit-learn
    • EDA
  3. OneStop AI logo

    Machine Learning & AI Intern

    OneStop AI · 1stop

    May 2025 – Aug 2025Remote

    • Improved model accuracy by 12% across 3 production AI projects through hyperparameter tuning and architecture changes.
    • Cut inference latency by 18% (220 ms to 180 ms) by refactoring preprocessing and applying model quantization.
    • Built end-to-end data pipelines for preprocessing, feature engineering and batched inference.
    • Added monitoring and model performance tracking with the senior engineers to make deployments more reliable.
    • PyTorch
    • asyncio
    • Data pipelines
    • Monitoring
View full résumé

Projects

4cooperating agents

AgentDesk

Multi-agent customer-support assistant

A planner agent routes each request to knowledge, data and action agents that use schema-validated tool calls. RAG over policy docs with citations, prompt-injection and PII guardrails, human approval for large refunds, and an eval harness that scores every change.

  • Python
  • FastAPI
  • Tool calling
  • RAG
  • SQL
  • Evals
25→92%extraction validity

DocGuard-VLM

Fine-tuned vision-language model

Qwen2-VL-2B fine-tuned with QLoRA for receipt field extraction and forgery detection. Scored on 120 held-out examples against the zero-shot model, field-extraction validity went from 25% to 92% and forgery F1 from 0.21 to 0.60.

  • PyTorch
  • PEFT / QLoRA
  • unsloth
  • Qwen2-VL
7orgs merged its fixes

CodeHound

Python static analyzer

AST checks for async-safety and correctness bugs: blocking calls in async code, fire-and-forget tasks, leaked file handles, mutable defaults. Its findings became merged fixes in unsloth, mem0, agno, Black, Weaviate, Xorbits and pydantic-ai. Zero dependencies, CI-ready, with a citable DOI.

  • Python
  • ast
  • pytest
  • GitHub Actions
−30%tokens per query

PathForge

AI placement-preparation platform

Generates role-specific study material and mock interviews with RAG over FAISS. A skill-gap scoring layer filters retrieval by the learner's level before prompting, so answers are personal instead of generic, and a roadmap engine builds a study plan from the gaps.

  • LangChain
  • OpenAI API
  • FAISS
  • FastAPI
  • React
3verification layers

MarkMe

Proxy-proof attendance system

Three checks before attendance counts: face recognition with liveness detection, GPS geofencing within 100 m, and a rotating 6-digit session key. Face detection runs on the device, so biometric data never leaves the phone.

  • Python
  • JavaScript
  • Face recognition
  • Geofencing
34→81%emotion accuracy

SoulSync

AI mental-health chatbot · Smart India Hackathon 2024

Generic sentence embeddings could not separate neutral text from distress (34% accuracy). Fine-tuning a sentence-transformer on mental-health data with multi-label emotion scoring raised it to 81%, and context tracking cut irrelevant replies by about 60%.

  • Python
  • Transformers
  • NLP
  • FastAPI

Open Source

50 pull requests merged across 30 organisations with 400k+ combined stars. Mostly real bug fixes (event-loop blocking, leaked connections, shared mutable state), each with a regression test and reviewed by the project's maintainers.

  • unsloth77k★
  • mem066k★
  • agno42k★
  • Black42k★
  • JAX (Google)36.6k★
  • pydantic29k★
  • marimo23k★
  • HuggingFace22k★
  • Snipe-IT15k★
  • Prowler15k★
  • redis-py13.6k★
  • OGX8.5k★
  • asyncpg8.1k★
  • scapy12.6k★
  • TagStudio7.3k★
  • Xorbits9.6k★
  • Celery3.1k★
  • pymodbus2.8k★
  • Galaxy1.9k★
  • AMD GAIA1.6k★
  • pyroute21.1k★
  • pyfakefsmerged
  • Microsoft (semantic-link-labs)merged
  • Sklearn-genetic-optmerged
  • lokymerged
  • fakeredismerged
  • Weaviatemerged
  • Apache Mavenmerged
  • Mercarimerged
  • palinodemerged

Highlights

All 50 merged pull requests
5
research works
1
accepted at IEEE
3
as first author
  • AcceptedIEEE ISED 2026 · NIT Warangal

    Machine Learning-Based Student Performance Prediction: A Comparative Study Using Ensemble Methods and Explainable AI

    Second author of six, with D. Jaggupalli and C. Mahanty

    Compares ensemble models for predicting student performance and uses SHAP so the drivers behind each prediction can be inspected rather than trusted blindly.

    • Ensemble methods
    • Explainable AI
    • SHAP
  • Abstract submittedSpringer book chapter · Lightweight Deep Learning for Efficient AI

    From Compression Ratios to Wall-Clock Gains: A Systems-Level Review of Lightweight Deep Learning on Edge Hardware

    First and corresponding author

    Argues that the gap between reported model compression and real latency or energy gains is a systems problem, and proposes a hardware-aware evaluation protocol.

    • Model compression
    • Edge AI
    • Quantisation
  • Abstract acceptedICTIRL-2026 · Springer Nature – Atlantis Press proceedings

    AI-Assisted Bug Detection and the Accountability Gap: A Case Study of an Open-Source Static Analyzer

    First author with D. Jaggupalli and R. Edupuganti

    Uses CodeHound's documented validation record to argue that AI developer tools should be judged on how their rules were validated, not only on whether an output was right.

    • Algorithmic accountability
    • Static analysis
    • AI governance
  • PreprintResearchGate · May 2026

    HAPS: A Hybrid AI Proctoring System for Unified Online and Offline Examination Integrity

    First author with D. Jaggupalli and J. Pujith

    Combines dual-stream CNNs, YOLO object detection and multi-modal behavioural signals to catch malpractice in both online and in-hall exams.

    • Computer vision
    • YOLO
    • Multi-modal
  • PreprintResearchGate · March 2026

    AI, ML and DL-Based Integrated Drone Detection and Autonomous Defence Systems: A Review

    Co-author with D. Jaggupalli and J. Pujith

    Surveys deep-learning architectures and deployment strategies for detecting drones and autonomous defence.

    • Deep learning
    • Detection
    • Survey

Achievements & Certifications

2025

Top 5 · AI & ML Hackathon

BITS Pilani Hyderabad · Techgyan

Placed in the top five at a 24-hour in-person AI/ML hackathon hosted at BITS Pilani, Hyderabad.

2026

Paper accepted at IEEE ISED 2026

NIT Warangal

Ensemble ML with SHAP explainability for student performance prediction, accepted for the IEEE conference proceedings.

2026

29 PRs merged in 18 organisations

Open source

Reviewed and merged by maintainers at HuggingFace, pydantic, Black, redis-py, Celery, unsloth, Apache and more, with code shipping on PyPI.

2024

Smart India Hackathon

Government of India

Built SoulSync, an AI mental-health chatbot, for India's largest national hackathon.

2025

212 LeetCode problems

LeetCode

91 easy, 108 medium and 13 hard over 127 active days, with an 82-day best streak.

2026

GitHub Galaxy Brain

GitHub achievements

Earned for answers accepted by the people who asked them, including one in Snipe-IT that became a merged fix.

Certifications