Skip to content

●  Lead AI Engineer — Eject Solutions


01  AI/ML Engineer02  Agentic AI & LLM Systems03  Business Analyst

I design agentic AI and LLM systems that leave the notebook, ship to production, and move the numbers a business actually reports on.

ScrollStart a conversation

01About

Engineering that earns its seat in the room.

Three years of turning machine learning into operating leverage — from data pipelines to autonomous agents that do real work.

I sit where engineering meets the boardroom. I build ML, NLP and agentic LLM systems end-to-end, then work alongside C-suite leaders to decide where those systems should point next.

I independently take AI work from idea to patent- and publication-ready — the same rigor goes into production pipelines as into peer-reviewed research.

Years building ML, NLP & agentic systems
3+
Business impact facilitated
₹1.4Cr+
Qualified leads generated
3,000+
Staff time automated weekly, across 4 organizations
20+ hrs

02Experience

From freelancer to lead, in under two years.

  1. Mar 2024 — Present

    Lead AI Engineer

    Eject Solutions Pvt Ltd

    Promotion path

    1. Freelance AI & Data EngineerMar 2024
    2. Business Analyst & AI Engineer
    3. Lead AI EngineerNov 2025
    • 01Built LLM-powered automation pipelines that replaced repetitive manual workflows across client operations.
    • 02Designed an agentic AI architecture that was adopted as the company's production blueprint for new builds.
    • 03Made model and pipeline deployment 40% faster.
    • 04Delivered INR 4L in cost reduction on an engagement with KL University.
    • 05Ran the data side of a lead-generation campaign that produced 3,000+ qualified leads.
    • 06Reduced sprint delays by 35%.
    • 07Shipped Power BI dashboards that gave leadership a live read on pipeline and performance metrics.

  2. Mar 2023 — Mar 2024

    Marketing Analytics Analyst

    INT360 Design Studio

    • 01Automated recurring reporting and data workflows, freeing the team from manual spreadsheet work.
    • 02Built campaign and performance dashboards used for weekly marketing decisions.
    • 03Designed and analysed A/B tests to validate creative and channel choices with evidence.
    • 04Ran cohort analyses to surface retention and engagement patterns across customer segments.

03Selected work

Ideas taken all the way — to production, patent, and print.

Five projects spanning production agent pipelines, a patent-bound robot, and peer-reviewed machine learning.
  1. 01 / 05Self-directed R&D

    Cybercrime rapid-response & intelligence platform

    MITHRA

    A self-directed R&D platform for responding to cybercrime quickly and turning incident signals into usable intelligence.

    • Engineered with Claude Code- and Codex-assisted development workflows.
    • Manuscript in preparation, pursuing publication in a Scopus-indexed journal.
    • Agentic AI
    • LLMs
    • Security
    • R&D
    Project pagefor MITHRA
  2. 02 / 05

    Force-adaptive, lizard-inspired tree-climbing harvester

    Aglier

    A robot that climbs trees and harvests fruit, deciding through a tiered edge–cloud architecture with an asynchronous narrative layer on top.

    • Sole inventor, drafting the patent specification.
    • Carries an onboard GBNI subsystem for state-aware reasoning.
    • Annotated patent drawing of the Aglier tree-climbing fruit-harvesting robot clamped to a trunk with four clamps, showing the AI camera module, articulated arm, gripper, collection basket, battery pack and communication module.
    • Timing diagram of the per-fruit harvest cycle in five stages: climb and reposition, vision confirmation and AI decision, manipulator approach, cut or grasp, and transfer to the basket, with a structured event log feeding the GBNI narrative tier.
    • Geometry of a claw pair closing around a tree trunk, showing claw length L, trunk radius r and closure angle theta, with the relationships r = L sin(theta/2) and theta = 2 arcsin(r/L).
    • Six alternative bark-engagement designs: claw friction clamp, microspine array, hybrid adhesive and microspine foot, tendon-driven wrap gripper, wheeled spring-cinch mechanism and pneumatic flexible clamp.
    • Robotics
    • Computer Vision
    • Embedded AI
  3. 03 / 05SSRN · 7531958

    State-aware multimodal agents for adaptive brand communication

    GBNI (V2) & Autonomous Agency Workflows

    Generative Brand Narrative Intelligence — an agent framework that reads context and emotional state to adapt how a brand speaks.

    • GBNI v2 on SSRN (Abstract ID: 7531958): consistency governance, causal self-optimisation and multi-tenant support on top of the EABN framework.
    • Builds on the original GBNI paper on SSRN (Abstract ID: 6845958).
    • Expanding into a 120+ page IEEE paper and a full book manuscript.
    • 5 autonomous agent pipelines in production, recovering 20+ hours every week.
    • Multi-agent
    • Multimodal
    • LLMs
    • Production
    Project pagefor GBNI (V2) & Autonomous Agency WorkflowsRead the write-upabout GBNI (V2) & Autonomous Agency Workflows
  4. 04 / 05₹1.4Cr+ facilitated
    Lead pipeline: Meta lead ads, Google Ads, LinkedIn and website, CRM and event forms feed 1 Extract (connectors.py, raw_leads), 2 Clean (Python, clean_leads), 3 Dedupe (Python and SQL, golden_leads), 4 Score (SQL, lead_scores), then 5 Report (campaign_performance).Meta adsGoogle AdsLinkedInWeb · CRMExtractconnectors.pyraw_leads0 rowsCleanPythonclean_leads0 rowsDedupePython + SQLgolden_leads0 rowsScoreSQLlead_scores0 rowsReportcampaign_performanceLeads by score tierHotWarmCold12345
    Live simulation · illustrative counts

    Four ad channels in, one scored lead list out

    Lead Generation Data Pipeline

    A Python + SQL pipeline that pulls leads from every campaign channel, cleans and de-duplicates them into one golden record per person, scores them, and reports which campaigns bring leads worth calling.

    • One connector layer (connectors.py) lands Meta lead ads, Google Ads, LinkedIn and website, CRM and event forms in a single raw table.
    • Python cleaning plus Python + SQL de-duplication build one golden record per lead before SQL scoring.
    • Scores roll up into a campaign_performance report that compares channels on lead quality, not just volume.
    • 3,000+ qualified leads generated and ₹1.4Cr+ in revenue facilitated.
    • Python
    • SQL
    • Data Engineering
    • Lead Scoring
    Read the write-upabout Lead Generation Data Pipeline

04Marketing

The growth side.

Growth marketing, business development and campaign analytics: the same data and automation skills, pointed at leads, admissions and elections. Toggle through the marketing side of my résumé.

  • ₹1.4Cr+

    Revenue facilitated

  • 3,000+

    Qualified leads

  • 29/29

    Election seats won

  • 20%

    Lower cost per lead

  • 30%

    Conversion lift

Growth and business development professional with 3+ years of experience in lead generation, campaign management and marketing analytics, backed by engineering skills to automate the work. I've delivered measurable results across online and offline campaigns: led the campaign for a panel that won all 29 of 29 election seats in an association election, achieved the top admission rate in a university admissions drive, produced 3,000+ qualified leads that facilitated INR 1.4 crore+ in revenue, and cut cost per lead by 20% through Meta A/B testing. I communicate with leadership, clients and students in English, Telugu and Hindi.

05Publications

On the record.

Peer-reviewed and preprint work. Three papers are on SSRN: a theory of marketing advantage under Algorithmic Parity and both GBNI papers (v1 and v2); a GBNI IEEE paper, a MITHRA Scopus-track manuscript and a book are in preparation.
  1. [1]

    Working paper · Marketing strategy

    “When Everyone has AI: A Theory of Marketing Advantage under Algorithmic Parity.”

    SSRN · Oct 2026 · Abstract ID: 7575263

    View source(opens in a new tab)
  2. [2]

    Working paper · GBNI v2

    “Generative Brand Narrative Intelligence v2: Consistency-Governed, Causally Self-Optimizing, and Multi-Tenant Extension of the Emotionally Adaptive Brand Narrative (EABN) Framework.”

    SSRN · Sept 2026 · Abstract ID: 7531958

    View source(opens in a new tab)
  3. [3]

    Working paper · GBNI v1

    “Generative Brand Narrative Intelligence: A State-aware Multimodal Agent Framework for Autonomous Emotionally Adaptive Brand Communication.”

    SSRN · May 2026 · Abstract ID: 6845958

    View source(opens in a new tab)
  4. [4]

    Peer-reviewed journal

    “Diabetes Data Analysis and Machine Learning Based Prediction Model on Streamlit Web App.”

    International Journal of Scientific Research and Engineering Development (ijsred.com) · Oct 2022

    View source(opens in a new tab)

06Capabilities

A toolkit, organised.

Six disciplines, one practice. Pick a category to see what sits inside it.

Agentic AI & LLMs

Autonomous, tool-using systems that run in production.

  • 01Multi-agent orchestration
  • 02LLM automation pipelines
  • 03Prompt engineering
  • 04Claude Code & Codex workflows

07Education

B.Tech, Data Science

KL University · 2019 — 2023

8.1CGPA

Relevant coursework

  • Machine Learning
  • Deep Learning
  • Natural Language Processing
  • Statistics & Probability
  • Data Structures & Algorithms
  • Database Systems

08Writing

Notes from the work.

Longer write-ups on the systems, research and decisions behind the numbers above.

  1. · 12 min

    When everyone has AI: where marketing advantage goes

    If every competitor rents the same AI from the same platforms, advantage has to come from somewhere else. My new working paper names the condition, Algorithmic Parity, and the five assets that stay scarce, Differentiation Capital.

    Read

  2. · 2 min

    What I'm building in 2026: agentic AI, a harvesting robot and three manuscripts

    A map of my current work: production AI agents, the GBNI framework, the Aglier fruit-harvesting robot, the MITHRA cybercrime platform, and the papers behind them.

    Read

  3. · 3 min

    What a business analyst brings to an AI team

    Why effective AI work starts with a business decision, not a model, and how A/B testing, cohort analysis and dashboards make AI projects succeed.

    Read

All writing

09Contact

Let’s build
something that thinks.

Open to AI engineering roles, agentic-systems builds and research collaborations. The fastest way to reach me is email.

Email me LinkedIn (opens in a new tab)
Email
adithyareddy639@gmail.com
Phone
+91-87125 69303
Based
Hyderabad, India

10Play

Before you go, play.

A small Pac-Man, built from scratch for this site. Keyboard, swipe or the on-screen buttons.