# Joseph Avula > Serving people better with AI behind the work. Source: https://javula.me/ (built 2026-11-12). Everything on javula.me, plus Joseph's profile, as plain Markdown. ## Quick facts - Name: Joseph Avula - Based in: Dallas, TX - Focus: Finance, Sales, People. Graduating December 2026. - Now: GTM Operator at Daybreak AI (May 2026 to now). Reports to the Chief Growth Officer. - Education: Master of Science in Supply Chain Management, UT Dallas (expected December 2026); Bachelor of Technology in Civil Engineering, GITAM (August 2024). - Licensing: FINRA Securities Industry Essentials (SIE) exam scheduled for November 19, 2026. - Strengths: executive outreach, client relationships, deal pipelines, and AI automation. - Languages: English, Telugu. - Contact: josephavula9@gmail.com · https://www.linkedin.com/in/josephavula/ ## Summary Business operator and people person: executive outreach, client relationships and deal pipelines, multiplied by AI fluency. At Daybreak AI he hosted a summit for 42 VP and C-suite leaders that produced 15 meetings and $4M in pipeline, and booked 8+ executive meetings through cold outreach. As an advisor to Ahuva, an Indian critical-infrastructure integrator, he helped win a $10M+ AI traffic-management contract. He also builds tools that read SEC filings and public data and run on a schedule. Each project below lists what it shows and its numbers. ## Experience ### GTM Operator, Daybreak AI Dallas, TX (hybrid) · May 2026 to present · Reports to the Chief Growth Officer. - Hosted a Nashville summit for 42 VP and C-suite leaders and briefed leadership on every guest. It produced 15 meetings, $4M in pipeline and 2 deals near close. - Booked 8+ executive meetings through cold outreach, including the CEO of a $600M auto-parts maker and a global-brand director (NDA signed). Translated CFO priorities for the product team on semiconductor pursuits. - Audited a $17.8M sales pipeline with AI tools ($15M had no close date), built a 4,266-company target account list, and shipped a proposal tool adopted company-wide. ### Advisor, Ahuva Visakhapatnam, India · 2024 to present - Advises the leadership of Ahuva, an Indian critical-infrastructure integrator, on bid strategy. - Met about 20 port officials, OEM partners and ministry leaders alongside leadership. Helped win a $10M+ (INR 100 Cr) AI traffic-management contract. - Sized India's 5-year market for electronic tolling and traffic-management contracts for leadership planning. ### Intern, Fusion AI Agent Development, Intellinum Inc. Dallas, TX · March 2026 to May 2026 - Built Oracle Fusion AI agents that automate freight-invoice validation and warehouse order checks. Shipped 5 production features. ### Summer Intern, Procurement, AR Cube Loyalty Manchester, UK · April 2024 to June 2024 - Ran vendor due diligence for an airline client, narrowing 80 candidates to 12 for the Phase 1 rollout. - Negotiated a $100K annual contract with an airline loyalty program's director and onboarded the airline as a client. ### Founder and Owner, Ophir Textiles & Jewelry India · November 2023 to March 2024 - Built a direct-to-consumer brand to $10K revenue in 3 months ($130 average order, 400+ items, 13 suppliers). ## Education - Master of Science in Supply Chain Management, The University of Texas at Dallas, Naveen Jindal School of Management (JSOM), Dallas, TX. Expected December 2026. - Bachelor of Technology in Civil Engineering, Gandhi Institute of Technology and Management (GITAM), Visakhapatnam, India. August 2024. ## Licenses, skills and more - Licensing: FINRA Securities Industry Essentials (SIE) exam scheduled for November 19, 2026. - Skills: Excel, CRM, AI automation (Claude Code). - Languages: English, Telugu. - Memberships: ISM DFW; ASCM North Texas. - Community: Volunteer, Countryside Bible Church (Southlake, TX). ## Projects Each project's sections are quoted from its page on javula.me, in Joseph's own voice. ### Window Board - Page: https://javula.me/projects/window-board/ - Area: Finance - Status: Live, updates weekly - What it is: For public companies in 17 U.S. metros, shows when each insider trading window opens and closes, plus the last day to adopt a 10b5-1 plan. Backtested on 1,323 real plan adoptions. - Key number: 80.9% of 1,323 real plan adoptions fell inside the predicted windows. A random day: 39.1%. - What it shows about Joseph: Turns SEC filings into a timing tool for advisors who work with executives. Backtest: 80.9% of 1,323 real 10b5-1 plan adoptions fell inside the predicted windows. A random day lands inside 39.1% of the time. #### The problem Executives can only trade their company's stock in set windows. Each company's insider trading policy says when a window opens and closes. Advisors who work with executives need those dates. They also need the last day to set up a 10b5-1 plan, and the first day that plan can sell. Every policy words its rule a different way. Reading them by hand, company by company, takes hours. #### What it does - Covers 2,489 public companies based in 17 U.S. metros. - For each one, it shows when the trading window opens and closes. - It shows the last day to adopt a 10b5-1 plan, and the first day that plan could sell. - Every date is an estimate. A company can close a window at any time, and insiders still need pre-clearance. #### How it works - It reads each company's insider trading policy (Exhibit 19 to its 10-K) and keeps the exact quote. - When a policy gives no date, it uses the most common rule and marks the date "typical rule." - Earnings dates come from 8-K filings. Future ones are predicted from last year's same quarter. - The 10b5-1 cooling-off period follows the SEC rule: 90 days after adoption, or 2 business days after the quarter's 10-Q or 10-K if that is later, never more than 120 days. - It rebuilds itself on GitHub Actions: new filings every Monday, and a full rebuild of companies and policies every quarter. Any failed check stops the run. - No person names, no holdings, no addresses. #### Proof - Backtest on 1,323 real 10b5-1 plan adoptions by officers and directors, from 10-Q and 10-K filings (as of November 12, 2026). - Each window was predicted using only earnings dates known before that quarter began. - 80.9% of adoptions (1,070 of 1,323) fell inside the predicted window. A random trading day lands inside 39.1% of the time. - With actual earnings dates: 82.6%. Form 144 plan dates: 81.1% of 3,203. - Policies the board read itself hit 84.0% (of 1,131). The typical-rule fallback hits 62.5% (of 192). #### Status Live at this address. It rebuilds itself every week from new SEC filings. #### Links - [SEC Rule 10b5-1](https://www.ecfr.gov/current/title-17/chapter-II/part-240/subpart-A/subject-group-ECFR71e2d22647918b0/section-240.10b5-1) ### Liquidity Radar - Page: https://javula.me/projects/liquidity-radar/ - Area: Finance - Status: Live - What it is: Dallas-Fort Worth executives about to get paid in pending buyouts, read from SEC merger filings, with the deal's bankers and lawyers. - Key number: Arcosa backtest: the proxy pointed to $166.7M for insiders 81 days before close. $184.0M at close. - Latest: As of October 31, 2026: six DFW companies are being sold. Filings show up to $288M coming to 66 insiders at five of them. The sixth is a $2.6B private sale, not in that total. - What it shows about Joseph: Finds liquidity events before they happen, which is how private wealth teams find new clients. Backtest on a closed deal (Arcosa): the merger proxy, filed 81 days before close, pointed to $166.7M for 13 insiders. $184.0M was paid at close. #### The problem Private wealth teams find new clients through liquidity events, like an executive's payout when a company is sold. SEC filings show those payouts before the deal closes. #### How it works - Finds pending buyouts of DFW companies in SEC filings, by the company's ZIP code. Private sellers are found through the buyer's F-4 or S-4 filing and a web search. - Reads what each insider gets from the merger proxy's golden parachute table. When there is no merger proxy yet, it uses the annual proxy's change-in-control table plus shares owned, and says so. - Lists each deal's bankers and lawyers from the filings. - Every number links to the filing it came from. Photos come from company and firm sites. - Backtest on a closed deal (Arcosa, Dallas): the proxy, filed 81 days before close, showed $166.7M for 13 insiders. They got $184.0M. #### Status Live, last updated October 31, 2026. The page names real executives from public filings, so it is kept out of search indexes (noindex) but open to anyone with the link. ### Hometown Stock Fates - Page: https://javula.me/projects/stock-fates/ - Area: Finance - Status: In progress - What it is: Every public company based in 17 U.S. metros in 2006, and what happened to it since: still trading, acquired, private, bankrupt or delisted. In progress. - Key number: 2,813 public companies in the 17 metros in 2006, out of 4,791 listed nationwide. - What it shows about Joseph: Builds a large, checkable data set from raw SEC filings: 4,791 listed U.S. companies nationwide in 2006, 2,813 of them in the 17 metros, a 30-company hand-checked answer key, and J.P. Morgan's 44% catastrophic-loss benchmark checked against its source. Not finished yet. #### The problem Executives and early employees often hold a lot of one stock. Advisors explain that risk with national numbers. J.P. Morgan found that 44% of all companies ever in the Russell 3000 from 1980 to 2020 had a catastrophic loss: a fall of 70% or more from peak that was not recovered. This project asks the same question for one city at a time. Of the public companies based in your city in 2006, what happened to them? #### What it does - For each of 17 U.S. metros, it lists every public operating company based there in 2006. - It gives each one a fate today: still trading, acquired, went private, bankrupt or delisted. - Each company that stopped trading links to the SEC filing that ended its run. - The headline is the metro's catastrophic-loss rate, shown next to J.P. Morgan's national figure. - Every page says it plainly: base rates, not a view on any company. #### How it works - Starting set: every company that filed an annual report (10-K) with the SEC in 2006 and was listed on NYSE, Nasdaq or AMEX. Headquarters comes from the ZIP code in that filing. - Funds, trusts and blank-check shells are dropped. Every drop is logged with a reason. - Fates come from each company's later SEC filings: a bankruptcy 8-K, merger filings followed by a delisting form, or recent annual and quarterly reports. - Survivors are priced with monthly closing prices since January 2006. Price return only. - Losses are bankrupt plus delisted plus survivors that fell 70% or more from peak and stayed down. Acquired and private companies never count as losses, so the rate is a floor. Each rate shows its sample size and a 95% interval. - Planned refresh: every quarter, by GitHub Actions. #### Proof so far - 4,791 listed U.S. companies nationwide in the 2006 starting set, before the metro filter. 2,813 of them are in the 17 metros. 7,234 other filers dropped, each with a reason. - A 30-company answer key, checked by hand against SEC filings, to test the fate rules. - 8 of 8 spot-check companies landed in the right metro. - J.P. Morgan's 44% benchmark quoted from its own PDF and checked line by line. #### Status In progress. The starting set, the price tool, the benchmark and the answer key are built (November 2026). The fate rules, metro pages and PDFs come next. Nothing is published yet, so there is no live link. ### Job Radar - Page: https://javula.me/projects/job-radar/ - Area: Tools - Status: Live - What it is: Watches 10 elite private wealth firms' own career sites every hour and emails each new fitting role with pay, a fit note and the apply link. Live on GitHub Actions. - Key number: 119 open roles seeded. Runs hourly on GitHub Actions. - What it shows about Joseph: Builds and runs unattended automation: 10 different career-site systems read every hour on GitHub Actions, with health checks and alerts. Seeded with 119 open roles. Live since November 7, 2026. #### The problem I am targeting private wealth management. The firms I track post roles on their own career sites. Each site runs on a different system. Checking 10 of them by hand every day is slow, and easy to skip. #### What it does - Every hour, it pulls every open job from 10 elite private wealth firms' own career sites. - It keeps roles that fit by title, city and years of experience. - Each new fit arrives as one email: pay with its source, a short fit note, and the apply link. - Each role also lands in a Google Sheet, where I track its status. #### How it works - Sources: each firm's own job feed, no login. They run on Workday, Greenhouse, BrassRing, ADP, Phenom, a GraphQL API and an Atom feed. - Schedule: GitHub Actions, every hour. It runs while my laptop is off. - It pulls every job, not a keyword search, and checks the count against the total each feed reports. - Filters: a title rule, a city rule, then the job's own page for pay and years required. - Pay: the posted range first, then the firm's pay field, then a Department of Labor wage filing for a similar title in the same metro, shown as an estimate. - Fit notes are written by Claude against my profile. - Health checks: a feed counts as broken if it errors, returns zero jobs, misses its own count, or falls below half of last week. Six broken runs in a row send an alert. A heartbeat email goes out every Monday. #### Proof - Live since November 7, 2026. - Seeded with 119 open roles across the 10 firms, sent as one summary. - First scheduled run: all 10 feeds healthy, 0 failed sends. #### Status Live. Runs every hour. The code is private, so there is no public link. ### Tender Radar - Page: https://javula.me/projects/tender-radar/ - Area: Ahuva - Status: Live - What it is: Reads public Indian government tender pages every morning at 7:30 a.m. CT and emails anything new that fits Ahuva, with the days left to bid. - Key number: Runs daily at 7:30 a.m. CT. - What it shows about Joseph: Turns scattered public tender sites into one daily email for a real business (Ahuva). Runs every day at 7:30 a.m. CT across 8 public buyers. #### The problem Ahuva, an Indian critical-infrastructure integrator I advise on bid strategy, bids on public infrastructure work. Tenders are posted on many separate government sites, each with its own layout. A missed posting is a missed bid. #### What it does - Every morning it reads public tender pages from 8 buyers: major ports, national highway bodies and a smart city. - It keeps tenders that match Ahuva's work, using a keyword list. - It emails anything new, with the days left to bid. #### How it works - One Python script with no outside packages. Each source is one small function, so adding a buyer means writing one function. - Schedule: every day at 7:30 a.m. CT. - It remembers what it already sent, so each tender arrives once. It tracks failure streaks for each source. - Email goes through a small Google Apps Script mailer that sends only to me. Job Radar uses the same mailer. - When a site posts no deadline, it estimates one as the posted date plus 45 days, and says so. #### Proof and coverage - Live since October 2026. Runs every day. - Covers 8 public buyers. - Skipped on purpose: portals behind a captcha. Some sites also do not answer from outside India. These gaps are written down, not hidden. #### Status Live. The code is private. Ahuva's site is public. #### Links - [ahuva.in](https://www.ahuva.in/) ### Ahuva Search Visibility - Page: https://javula.me/projects/ahuva-search-visibility/ - Area: Ahuva - Status: Live - What it is: Working to make Ahuva the company Google and AI assistants name for its buyers' searches: buyer queries scored on 5 engines, and sourced pages on ahuva.in. - Key number: 65 buyer queries in the baseline, across Google, ChatGPT, Gemini, Claude and Perplexity. - What it shows about Joseph: Runs a measured search and AI-answer program for a real company: 65 buyer queries scored on 5 engines. Baseline (October 29, 2026): 24 of 650 possible points. Re-test due on or after December 24, 2026. #### The problem Ahuva is an Indian critical-infrastructure integrator I advise on bid strategy. The people who buy its work, and the consultants who write tender specs, now ask Google and AI assistants who can do the job. If Ahuva is not in those answers, it is not on the shortlist. #### What it does - Tracks 65 buyer queries on Google, ChatGPT, Gemini, Claude and Perplexity. - Fixes ahuva.in so search engines and AI tools can read and cite it. - Publishes sourced reference pages on the technical questions buyers ask. #### How it works - Queries come from government plans, dated tenders and live Google autocomplete. - Scoring, for each query on each engine: 2 means it cites ahuva.in (Google: top 3). 1 means it names Ahuva (Google: top 20 or the AI Overview). 0 means Ahuva is absent. - Site fixes so far: sitemap fixes, IndexNow pings so new pages get crawled fast, and four reference articles. - A weekly loop finds new topics from Google autocomplete and competitor sitemaps, and measures results. - A monthly build writes up to 3 new pages. Every fact is logged with its source link and quote. A separate check verifies each page before it ships. - Each new page comes with predictions, scored at day 28 and day 56. #### Proof - Baseline, October 29, 2026: 24 of 650 possible points. That is the starting line. - Perplexity already cited ahuva.in 7 times at baseline. - Most Google, ChatGPT and Gemini runs failed during the baseline, so those scores are partial. Claude and Perplexity are complete. - Re-test due on or after December 24, 2026. #### Status Live. The weekly and monthly loops are scheduled. #### Links - [ahuva.in](https://www.ahuva.in/) ## Sales work at Daybreak AI Tools Joseph built in his current job at Daybreak AI. Listed on the Projects page; no pages of their own. - Sales skills for Claude: Nine Claude Code skills for sellers: RFP answers, business cases, call debriefs and account research. An independent AI judge graded each one. - CFO signal list: Found CFOs naming the problem on earnings calls. Narrowed 515 records to 39 qualified leads, each with a source. - Pipeline inspector: Reviewed 55 open deals in 23 minutes. Every deal got a verdict and a dated reason. - Email finder in Slack: Type a name and company, get a verified work email in seconds. Built in one day. - Reply desk: Owned replies for a campaign to 3,470 executives. Cut reply time from days to about an hour. ## Site pages ### Home (https://javula.me/) One screen: the name (Joseph Avula), the one-liner "Serving people better with AI behind the work.", and two buttons, Projects and Story. Under them, two live usage counters. Claude Code: about 10.1 billion tokens used since May 2026 (10,098,535,596 on 2026-11-11, then rising with the clock at his average rate; an estimate from usage logs). Wispr Flow: 682,545 characters dictated since April 2026 (same method, from the Wispr Flow app). Footer: LinkedIn and email. The page also holds a short text profile for screen readers and AI tools, not shown on screen: - Based in Dallas, TX. Focus: Finance, Sales, People. Graduating December 2026. - GTM Operator at Daybreak AI (May 2026 to present). Hosted a summit for 42 VP and C-suite leaders that produced 15 meetings and $4M in pipeline. - Advisor to Ahuva, an Indian critical-infrastructure integrator, on bid strategy (2024 to present). Helped win a $10M+ AI traffic-management contract. - MS in Supply Chain Management, UT Dallas, December 2026. - FINRA SIE exam: November 19, 2026. - Links to Story, Projects and this file. ### Projects (https://javula.me/projects/) Things I built for finance, for sales, and for myself. A filter shows All, Finance, Sales or Miscellaneous. Ahuva: an Indian critical-infrastructure integrator I advise on bid strategy. - Window Board: Insider trading windows and 10b5-1 plan dates for public companies in 17 U.S. metros. 80.9% of 1,323 real plan adoptions fell inside the predicted windows. A random day: 39.1%. (Live, updates weekly) - Liquidity Radar: Dallas-Fort Worth executives about to get paid in pending buyouts, read from SEC merger filings. Arcosa backtest: the proxy pointed to $166.7M for insiders 81 days before close. $184.0M at close. - Hometown Stock Fates: Every public company based in a metro in 2006, and what happened to it since. 2,813 public companies in the 17 metros in 2006, out of 4,791 listed nationwide. (In progress) - Job Radar: Watches 10 private wealth firms' job feeds every hour and emails each new fitting role. 119 open roles seeded. Runs hourly on GitHub Actions. - Tender Radar: Reads Indian government tender pages every morning and emails anything new that fits Ahuva. Runs daily at 7:30 a.m. CT. - Ahuva Search Visibility: Working to make Ahuva the company search engines and AI assistants name for its buyers' searches. 65 buyer queries in the baseline, across Google, ChatGPT, Gemini, Claude and Perplexity. - Sales skills for Claude (Sales, Daybreak AI): Nine Claude Code skills for sellers: RFP answers, business cases, call debriefs and account research. An independent AI judge graded each one. - CFO signal list (Sales, Daybreak AI): Found CFOs naming the problem on earnings calls. Narrowed 515 records to 39 qualified leads, each with a source. - Pipeline inspector (Sales, Daybreak AI): Reviewed 55 open deals in 23 minutes. Every deal got a verdict and a dated reason. - Email finder in Slack (Sales, Daybreak AI): Type a name and company, get a verified work email in seconds. Built in one day. - Reply desk (Sales, Daybreak AI): Owned replies for a campaign to 3,470 executives. Cut reply time from days to about an hour. ### Story (https://javula.me/story/) Joseph's about-me text beside a slowly turning globe with the countries he grew up in. The résumé above (experience, education, licensing, projects) is in the page HTML for screen readers and AI tools, not shown on screen. #### About me I grew up in four countries, which taught me how to read people fast and earn their trust. I've been selling ever since middle school: candy, then gemstones, then national infrastructure, and now AI software. #### Principles Care about people, always. Understand their problems better than they do, then picture the best outcome and work toward it together. #### Vision Let AI carry the busywork so people get their time back for the work only people can do: trust, judgment and care. #### Hobbies Muay Thai, swimming and perfumes. #### Where I grew up - South Korea, Seoul: 2002–2005 - UAE: 2005–2012 - USA: 2012–2016 - India: 2017–2024 - USA: 2024–now